To the Editor The report by Gray et al1 on the association of anticholinergic use with incident dementia may have overestimated the risk of anticholinergic use due to residual confounding. Participants with the highest level of cumulative exposure had, at baseline, more than double the rate of stroke and triple the rate of fair or poor health by self-report than participants with no exposure. They also had higher rates of hypertension, diabetes mellitus, and obesity; were twice as likely to be women; and 4 times as likely to have high levels of depressive symptoms. Since these independent risks for dementia were present at baseline and were correlated with anticholinergic exposure, the anticholinergic-dementia association may reflect residual confounding and could persist despite lag-time analysis, which discounts exposure only in the 2 years prior to dementia onset. Stratified analyses limited to women or men who were free of these potential confounders at study entry may reduce error due to residual confounding and give more accurate estimates of effect size.
Fried R. Anticholinergic Use With Incident Dementia. JAMA Intern Med. 2015;175(9):1577. doi:10.1001/jamainternmed.2015.2565