On 2017 Aug 09, David Keller commented:

Two Large Epidemiological Studies Reach Opposite Conclusions Regarding Dementia Risk From PPI Use

A recent German study [1] found a significant 44% increased incidence of dementia with regular proton-pump inhibitor (PPI) antacid use, in a cohort of 73,679 baseline non-demented persons over age 75, of whom 29,510 developed dementia over 8 years, and 2950 were regular PPI users. The authors of that study boldly concluded that "avoidance of PPI medication may prevent the development of dementia" and called for a randomized controlled trial of PPIs to better assess this risk.

The current study [2] looked at the 70,718 cases of Alzheimer disease (AD) diagnosed in Finland over a 7-year period and found no difference in risk based on regular PPI use, and a confirmatory lack of association of risk with dose of PPI, or duration of PPI use.

Which of these two large, painstaking studies in similar populations is correct? PPI use either does or does not increase the risk of dementia, and it is crucial to know which is true. Millions take PPI medication daily to control acid reflux, but would be better off taking less effective or tolerable treatments if PPIs increase the risk of dementia.

Taipale and colleagues identified PPI users from medication-purchase data, whereas Gomm and colleagues identified PPI users from prescription data, which did not indicate whether the prescriptions were actually filled. Obviously, persons who purchase a medication are more likely to be taking it than persons who are merely prescribed it. However, the effect of non-compliance with prescribed PPI therapy in the Gomm study is to actually increase the risk of dementia caused by taking PPI medication. The more non-compliance among patients in Gomm's PPI group, the greater the dementia-increasing effect PPIs must have had on compliant patients to explain the observed overall increased dementia in the PPI group. So, the difference in how PPI-takers were identified cannot account for the discrepant results of these studies.

How much credence can we place in epidemiological results if two large studies can report polar opposite conclusions with such apparent certainty? Is it ethical, or even possible, to randomize patients to less-effective medications than PPIs in a controlled trial, to definitively determine whether PPIs truly increase the risk of dementia?

References

1: Gomm W, von Holt K, Thomé F, Broich K, Maier W, Fink A, Doblhammer G, Haenisch B. Association of Proton Pump Inhibitors With Risk of Dementia: A Pharmacoepidemiological Claims Data Analysis. JAMA Neurol. 2016 Apr;73(4):410-6. doi: 10.1001/jamaneurol.2015.4791. PubMed PMID: 26882076.

2: Taipale H, Tolppanen AM, Tiihonen M, Tanskanen A, Tiihonen J, Hartikainen S. No Association Between Proton Pump Inhibitor Use and Risk of Alzheimer's Disease. Am J Gastroenterol. 2017 Jul 11. doi: 10.1038/ajg.2017.196. [Epub ahead of print] PubMed PMID: 28695906.

On 2017 Aug 09, David Keller commented:

Two Large Epidemiological Studies Reach Opposite Conclusions Regarding Dementia Risk From PPI Use

A recent German study [1] found a significant 44% increased incidence of dementia with regular proton-pump inhibitor (PPI) antacid use, in a cohort of 73,679 baseline non-demented persons over age 75, of whom 29,510 developed dementia over 8 years, and 2950 were regular PPI users. The authors of that study boldly concluded that "avoidance of PPI medication may prevent the development of dementia" and called for a randomized controlled trial of PPIs to better assess this risk.

The current study [2] looked at the 70,718 cases of Alzheimer disease (AD) diagnosed in Finland over a 7-year period and found no difference in risk based on regular PPI use, and a confirmatory lack of association of risk with dose of PPI, or duration of PPI use.

Which of these two large, painstaking studies in similar populations is correct? PPI use either does or does not increase the risk of dementia, and it is crucial to know which is true. Millions take PPI medication daily to control acid reflux, but would be better off taking less effective or tolerable treatments if PPIs increase the risk of dementia.

Taipale and colleagues identified PPI users from medication-purchase data, whereas Gomm and colleagues identified PPI users from prescription data, which did not indicate whether the prescriptions were actually filled. Obviously, persons who purchase a medication are more likely to be taking it than persons who are merely prescribed it. However, the effect of non-compliance with prescribed PPI therapy in the Gomm study is to actually increase the risk of dementia caused by taking PPI medication. The more non-compliance among patients in Gomm's PPI group, the greater the dementia-increasing effect PPIs must have had on compliant patients to explain the observed overall increased dementia in the PPI group. So, the difference in how PPI-takers were identified cannot account for the discrepant results of these studies.

How much credence can we place in epidemiological results if two large studies can report polar opposite conclusions with such apparent certainty? Is it ethical, or even possible, to randomize patients to less-effective medications than PPIs in a controlled trial, to definitively determine whether PPIs truly increase the risk of dementia?

References

1: Gomm W, von Holt K, Thomé F, Broich K, Maier W, Fink A, Doblhammer G, Haenisch B. Association of Proton Pump Inhibitors With Risk of Dementia: A Pharmacoepidemiological Claims Data Analysis. JAMA Neurol. 2016 Apr;73(4):410-6. doi: 10.1001/jamaneurol.2015.4791. PubMed PMID: 26882076.

2: Taipale H, Tolppanen AM, Tiihonen M, Tanskanen A, Tiihonen J, Hartikainen S. No Association Between Proton Pump Inhibitor Use and Risk of Alzheimer's Disease. Am J Gastroenterol. 2017 Jul 11. doi: 10.1038/ajg.2017.196. [Epub ahead of print] PubMed PMID: 28695906.