Proton Pump Inhibitor Use and Risk of Dementia: An Umbrella Review of Systematic Reviews and Meta-Analyses

Clinical studies about the association between long-term PPI use and dementia have been inconsistent, with systematic reviews and meta-analyses (SR/MA) reaching disparate conclusions.

This umbrella review of SR/MA-or overview of reviews-attempts to clarify the current understanding of any possible association linking PPI use to dementia.

We searched EMBASE, MEDLINE, Cochrane Database of Systematic Reviews (CDSR), and ISI Web of Science for fully published SR/MA in English up to June 2024 using controlled vocabulary and text words [(1) proton pump inhibitor and (2) dementia (dementia and mild cognitive impairment, Alzheimer disease)]. SR/MA methodological overall confidence was scored using the AMSTAR-2 tool.

Eleven SR/MA published (May 2016-June 2024) described 21 studies (one randomized trial). The number of studies in these SR/MA ranged from 4 to 14. Six of the included SR/MA reported data on Alzheimer's disease. Only one of the high-overall confidence SR/MA significantly associated PPI use with dementia. None of the six SR/MA investigating PPI use and Alzheimer's disease showed a significant association. Significant associations with dementia in subgroup analyses were found in SR/MA for older patients (one of three SR/MA) (Hazard ratio (HR) = 1.39 (1.17;1.65)), long-term PPI use (one of four SR/MA) (HR = 1.28 (1.12; 1.46)), short-term PPI use (one of four SR/MA) (Risk ratio (RR) = 1.62 (1.40-1.86)), when only considering higher quality studies (one of one SR/MA) (RR = 1.54 (1.12; 2.13)), among European populations (one of three SR/MA) (HR = 1.46 (1.23-1.73)), and cohort studies (two of four SR/MA) (RR = 1.36 (1.19; 1.55)), (RR = 1.44 (1.36; 1.52)).

Most SRs/MAs suggest no link between PPI use and dementia. Given the low certainty of existing evidence, a true causal relationship appears unlikely.