Our psychiatrist-reviewer analyzes news about proton pump inhibitors & dementia risk

Guest post by Dr. Susan Molchan, a psychiatrist with extensive experience in clinical research at the National Institutes of Health. She is the former program director for clinical dementia research at the National Institute on Aging.  And she is one of our most active contributors.

heartburnThis week, many news outlets are reporting that so-called heartburn drugs – proton pump inhibitors (PPIs) such as omeprazole (Prilosec), esomeprazole (Nexium), and lansoprazole (Prevacid)…

….based on an observational study by German researchers and published in JAMA Neurology. The journal manuscript is entitled, “Association of Proton Pump Inhibitors with Risk of Dementia.” (italics mine)

To report the findings accurately, journalists need to watch their language. An association does not indicate that the drugs cause or increase the risk of dementia. This could only be shown by a true experiment, a randomized controlled trial, where one group of people is put on PPIs for a long time and one on a placebo. Many journalists noted this in their articles. A randomized controlled trial is not likely to be done in this case though, and editorial commentator and epidemiologist Dr. Lewis Kuller from the University of Pittsburgh noted that further studies can look at people who are diagnosed with dementia compared to those without dementia where there is a good history or PPI use (case-control studies), and can see if PPI use is associated with pathological changes in the brain.

The researchers themselves overstate their findings, in my opinion, saying in their abstract that those patients taking PPIs “had a significantly increased risk of incident dementia compared with the patients not receiving PPI medication…The avoidance of PPI medication may prevent the development of dementia.”

The researchers examined an insurance claims database that included over 73,000 people, 75 and older, and free of dementia at the beginning of the study. They found that over 7 years those who took PPIs had a 44% increased probability of receiving a diagnosis of dementia (reported as a hazard ration of 1.44)

As Dr. Kuller points out, the German researchers found a similar association in another long term study of primary care patients, and this enhances the validity of their finding. Dr. Kuller points out reasons the drugs could cause an increase in dementia: they cross the blood-brain barrier, they can lower vitamin B12 levels, which, if too low, causes dementia, and in animal studies PPIs interact with amyloid and tau, proteins that muddle the brain in Alzheimer’s disease. He also notes why PPIs may not be causal: older people take many drugs, and these drugs are taken for a number of conditions. He notes that polypharmacy, taking many drugs, was also noted to be a risk factor for dementia in both German studies. People who take lots of drugs for several medical conditions (that may themselves increase dementia risk) go to the doctor more, hence may tend to be diagnosed with dementia more.

(Publisher’s note:  It always befuddles me why journalists wouldn’t take advantage of the freebie 2nd opinion that journal editorials provide them.  Why wouldn’t you read Kuller’s editorial for another perspective on the work you’re reporting on?  Why would you go with a single-source story – practicing the stenography of regurgitating what was reported by the researchers – and NOT at least nod in the direction of the accompanying editorial?  Fortune did.  Forbes did. MedPageToday did. NPR did, but didn’t cite the questions we captured, above. Others didn’t acknowledge the editorial at all. ) 

Since PPIs debuted over 25 years ago, the Food and Drug Administration (FDA) has noted that the drugs are associated with increased risk of fractures in the hip, wrist, and spine; lower magnesium, calcium; and infection with Clostridium difficile, a form of severe diarrheal infection that can be life-threatening especially in older people.

Millions of people take these drugs for symptoms of acid reflux or heartburn, or gastro-esophageal reflux disease, a mouthful and marketing triumph popularly known as GERD. Many who take them do not need them. A study from the Johns Hopkins School of Medicine for example showed that over 60% of PPIs prescribed in an inpatient setting were not for an FDA approved or clinically indicated condition, and many of these prescriptions were continued once patients went home.

In the U.S., drugs that people will take long-term are heavily marketed to both doctors and patients. We don’t know anything about a drug’s long-term effects when it is initially approved. Careful observational such as that reported are important to fill out this information. Although it still wouldn’t be definitive, If future studies can show that people taking higher compared to lower doses or for longer versus shorter time periods have higher probabilities of dementia, this would add substantially to the “causal” side of the evidence stack.

Journalists, medical journal editors, and scientists themselves can help people better understand studies and the risks of the drugs they take if they would take a little more care to watch their language.  We have a primer that can help with suggestions for accurate language to use to describe the difference between correlation and causation.

(Also see the analysis by the NHS Choices Behind the Headlines site, “Link between indigestion drugs and dementia ‘inconclusive’ “)

You might also like

Comments (15)

Please note, comments are no longer published through this website. All previously made comments are still archived and available for viewing through select posts.

Iggy Brady

February 16, 2016 at 9:30 pm

This is wonderful thanks! Hazard ratios tend to overstate absolute effects.
Two quick quibbles about this otherwise clear article: randomized trials are not strictly needed to prove risk or causation. Falls onto an outstretched hand most assuredly causes distal radius fracture, alcohol causes cirrhosis, asbestosis causes pleural placques, etc. Treatment effects require randomization; causation comes from a wide variety of observations. Second point: the authors mention controlling for polypharmacy. I don’t know if they do it, or how, but they claim it. Thanks!

    Susan Molchan

    February 19, 2016 at 8:01 am

    Thanks Iggy Brady and good points–also an accumulation of observational studies w/ large hazard ratios, along with animal data can be taken as evidence of causation: cigarettes causing lung cancer for example. The German authors did control for polypharmacy but still very hard to tease apart as noted in the accompanying editorial in the journal.
    Phil Foster, I considered adding a discussion of the benzo data but wanted to focus. Interestingly a study in BMJ this month found little association between their use and dementia, contradicting prior studies. Egle I also considered adding all the other associations of problems w the PPIs, but didn’t again in the interest of focus.

Phil Foster

February 17, 2016 at 2:02 am

First, it’s benzodiazepines, now it’s PPI’s causing dementia. How about old age causing dementia. Research has shown that our brains shrink as we age and that is the cause of dementia, now where did I leave my coffee cup?


February 17, 2016 at 7:45 am

Aside from dementia risk, PPIs have been found to alter patients’ overall health in other negative ways not indicated in your article:

James starr

February 17, 2016 at 8:31 am

“However, the study doesn’t make clear whether PPIs are also available over the counter in Germany, as they are in the United States, said Dr. Arun Swaminath, director of the inflammatory bowel disease program at Lenox Hill Hospital in New York City.

If over-the-counter PPIs are available, then more people might have been taking them and the dementia risk described in this paper could be overestimated, Swaminath said.”


February 17, 2016 at 8:35 am

The article above is thoughtful and well reasoned: a cool sip of water in a desert of superficiality. The name of the game now days is attention, measured by the number of clicks that an article garners. Perhaps thoughtful and well reasoned articles should be written all in caps with a few expletives thrown in……………..NOT!

Mark Ketterer PhD

February 17, 2016 at 1:59 pm

One can imagine several spurious reasons why an association might be observed between PPI usage and dementia: More pills (an indicator of more chronic illnesses that can cause dementia) leads to more GI distress & PPI use; lower tolerance of GI discomfort (a collateral finding in Demented patients) leads to more “complaints” and more scrips; reduced efficacy of meds used to control chronic risk factors (HTN or DM) because of reduced absorption; etc. On the other hand, if PPIs do reduce b vitamin absorption or increase CNS amyloid, a plausible pathway may exist. But even if the 44% increase is real, it is a small effect. Many of us believe that effect sizes below 100-200% in this kind of study should be taken with a grain of salt.

Tim Church

February 18, 2016 at 12:09 pm

PPIs were originally labeled for shot-term use based on 14-day courses of treatment for acute symptoms of reflux. They’re now prescribed and consumed like candy on an on-going basis. Whether they do or do not cause dementia (and I seriously doubt they do) their widespread use seems to be unwisely uncontrolled and higher-quality follow-up studies need to be undertaken.

    Stephen Cox, MD

    February 23, 2016 at 3:28 pm

    Based on the initial clinical studies and experience the PPI’s were considered very “clean” meaning they had great results for Peptic Ulcers and Reflux with apparently minimal side effects. With years of use and overuse and additional studies more side effects emerged. By then they were otc. instead of rx only as in the beginning. The FDA should be allowed to reconsider drugs like this to make them rx only again. A better example are NSAIDs, like motrin(ibuprofen), which has significant morbidity and mortality even though it has benefits. Some medicines need appropriate physician monitoring to help minimize toxicity.

June fisher

February 20, 2016 at 2:01 am

As a person who has been taking lansoprazol for approximately 15 years l am very concerned about this recent news on dementia and other very serious implications. I have barretts oesophagus and have been told l need to take lansoprazol 30th daily for the rest of my life. I have over the last 2 years had this dosage reduced to 15 ml because l was constantly suffering from a gurgling stomach and told my dosage may be too high. My concern as a lay person is how are we to know what to do for the best when even medical people seem to have completely different opinions. Furthermore nobody has ever told me about lack of vitamin b 12 absorption. This information is frightening to say the least. We are real people with real concerns about our health and in my opinion we need to be better informed and advised. We are lay people not doctors. Please consider this

    May Williamson

    February 22, 2016 at 6:29 pm

    I have taken omeprazole for 25 years , have tried to wean myself off them several times with increased stomach pain. This medication , with long term use- absolutely causes B12 deficiency – I have it and have been getting injectible B12 for 10 years. However, untreated GERD can lead to esophageal cancer. What to do?!?


February 22, 2016 at 8:39 pm

I’ve been on nexium 40 mg ever sense it hit the market.it’s the only thing that helps my Gerd .I’ve tried to ween off but am miserable. What to do? Is there any other options.

    Gary Schwitzer

    February 23, 2016 at 8:17 am


    As our comments policy clearly states, we cannot and will not give medical advice on this site.

    Note to other readers: we are clamping down more strongly on our policy asking commenters to leave a full name – first and last. We published Darlene’s because we often get asked for personal advice and it’s worth reminding users periodically that we will not answer any such questions.

    Gary Schwitzer

Arthur Leibovitz

February 26, 2016 at 4:44 am

A main question should be : why those patients were treated with PPI drugs ? They probably had GERD, Peptic Ulcer or similar conditions all closely related to the presence of Helicobacter Pylory .
And Helicobacter Pylori is repeatedly reported in the last years as associated with Cognitive Decline !!

John H Peterson

February 28, 2016 at 10:31 am

http://www.slideshare.net/PAWANVKULKARNI/review-of-new-alerts-on-proton-pump-inhibitors-ppi-adverse-effects-2015-55022824 have a look at these slides…. its not just dementia… but list of side effects of ppi are increasing