PPIs and Dementia: Is Causality Really There?

April 15, 2016

"Keep calm and carry on" suggests a thoughtful gastroenterologist whose patients continue to ask for reassurance.

[[{"type":"media","view_mode":"media_crop","fid":"47734","attributes":{"alt":"","class":"media-image media-image-right","id":"media_crop_5402480770581","media_crop_h":"0","media_crop_image_style":"-1","media_crop_instance":"5648","media_crop_rotate":"0","media_crop_scale_h":"0","media_crop_scale_w":"0","media_crop_w":"0","media_crop_x":"0","media_crop_y":"0","style":"margin: 8px 10px; float: right; width: 287px; height: 215px;","title":"©HalfPoint/Shutterstock.com","typeof":"foaf:Image"}}]]Proton pump inhibitors (PPIs) are among the most widely used medications, both as prescriptions and purchased over-the-counter (OTC) and treat gastroesophageal reflux disease (GERD), dyspepsia, and peptic ulcer disease. Although their safety profile is favorable, PPIs are associated with known adverse events. These are rare, but include renal dysfunction, electrolyte imbalances, vitamin deficiencies, and increased risk of pneumonia and enteric infections. In a recent prospective multicenter study of elderly patients, (aged ≥75 years), German investigators examined the association between PPI use and dementia risk.1 The investigators found that PPIs were significantly associated with increased risk of dementia. The results, published in February in JAMA Neurology, have been widely publicized in both professional and lay media. I continue to get calls from patients or their relatives who are concerned about continuing to take their prescribed PPIs.

Are the results “true?” Should our older and elderly patients taking PPIs be concerned about a possible increase in cognitive decline? One more look at the data is useful.

The study data come from a database of the largest mandatory public medical insurer in Germany, which includes one-third of the overall population and up to 50% of the elderly population. Baseline data on demographics and comorbid conditions were extracted on the more than 73,000 patients. Follow-up assessments occurred every 18 months. At baseline, the cohort had no underlying dementia, which was defined by an ICD-10 (International Classification of Diseases) code documented at least twice within an 18-month interval in their medical records. Regarding exposure to PPI, the authors searched for at least 1 PPI prescription or refill per quarter for at least 18 months. Cases with occasional PPI use were excluded. The main outcome was development of dementia, death, or neither at the end of the study period.

[[{"type":"media","view_mode":"media_crop","fid":"47737","attributes":{"alt":"","class":"media-image media-image-right","id":"media_crop_5338164092188","media_crop_h":"0","media_crop_image_style":"-1","media_crop_instance":"5649","media_crop_rotate":"0","media_crop_scale_h":"0","media_crop_scale_w":"0","media_crop_w":"0","media_crop_x":"0","media_crop_y":"0","style":"height: 241px; width: 250px; margin: 8px; float: right;","title":"©Lightspring/Shutterstock.com ","typeof":"foaf:Image"}}]]During the study period, 29,510 cases of dementia were diagnosed (59% with at least two types of dementia, 31% unspecified, 6% vascular, and 3% Alzheimer). After controlling for confounding variables such as comorbid depression, diabetes, ischemic heart disease, stroke, and use of 5 or more medications, the authors found a 1.44 increase risk of dementia among patients prescribed PPI therapy (95% confidence interval 1.36-1.52). The risk increased with age and was highest in the aged 75-79 year-old cohort. There was no difference in dementia risk among the three mostly commonly prescribed PPIs: omeprazole, esomeprazole, and pantoprazole.

PPIs have been shown to reduce vitamin B12 levels and increase β-amyloid in mice, offering biologic plausibility for potential increase in cognitive decline. However, several observational studies demonstrating significant associations could no longer demonstrate such results when controlled studies were later performed. In other words, causality is difficult to establish in [[{"type":"media","view_mode":"media_crop","fid":"47742","attributes":{"alt":"","class":"media-image media-image-right","id":"media_crop_2996963884983","media_crop_h":"0","media_crop_image_style":"-1","media_crop_instance":"5650","media_crop_rotate":"0","media_crop_scale_h":"0","media_crop_scale_w":"0","media_crop_w":"0","media_crop_x":"0","media_crop_y":"0","style":"float: right; width: 171px; height: 200px; margin: 8px;","title":"©Africa Studio/Shutterstock.com ","typeof":"foaf:Image"}}]]such studies. For example, observational studies suggested that PPI may affect the metabolism of clopidogrel leading many cardiologists to discontinue the PPI. The topic generated an abundance of news, similar to this current study, as well as heated debates. A randomized placebo controlled study was performed to help better define the relationship.2 The Kaplan-Meier curves for development of ischemic heart disease, the primary end point of the study, were nearly identical. However, patients on clopidogrel and placebo had a significantly increased risk of gastrointestinal bleeding. Following this study, clinicians became less concerned about the interactions between PPI and clopidogrel. We know, however, that conducting randomized controlled trials for every scientific question is challenging, expensive, and essentially impractical.

Although the PPI and dementia study raises interesting and important questions, several limitations urge caution in our over-interpretation of the results and offer points with which to counsel concerned patients:

First: it is challenging to assess whether a patient takes a medication based on prescriptions and refills.

Second: it is difficult to control whether patients in the control group used non-prescribed OTC PPI.

Third: the investigators use of diagnostic codes rather than validated instruments to establish diagnosis is suboptimal. There was lack of adjustment for well-recognized risk factors such as socioeconomic status, family history, obesity, heavy alcohol use, hypertension, atherosclerosis, and smoking.  

And, last: the proposed mechanism of PPI harm is elevation of β-amyloid which is independently associated with Alzheimer’s disease, however less than 3% of the cohort had this diagnosis.

In general, I advise my patients to take the smallest effective clinical dose of a medication when indicated. If a medication is not indicated, then it should be discontinued and the fact is PPIs are overly prescribed.  With that said, in those who require PPI (eg, GERD symptoms limiting quality of life or a patient with recurrent peptic ulcer disease who uses antiplatelets), we should "keep calm and carry on."


1. Gomm W, von Holt K, Thomé F, et al. Association of proton pump inhibitors with risk of dementia: a pharmacoepidemiological claims data analysis. JAMA Neurol. 2016 Feb 15 [Epub ahead of print].

2. Bhatt DL, Cryer BL, Contant CF, et al. Clopidogrel with or without omeprazole in coronary artery disease. N Engl J Med. 2010;363:1909-1917.