To the Editor In their recent publication, Gomm et al1 found that regular proton pump inhibitor (PPI) use was associated with a 44% increased risk for dementia based on claims data from a large German health insurer. The authors concluded that avoidance of PPI use may prevent the development of dementia. In addition to the established uncertainty inherent to claims data analyses2 and limitations of inferring causation based on observational data, we believe the design of the multivariable analysis did not adequately account for confounding by medication use.