Increased risk of dementia among patients with overactive bladder treated with an anticholinergic medication compared to a beta-3 agonist: a population-based cohort study

Objective To determine if there is an increased risk of dementia among patients with overactive bladder (OAB) starting an anticholinergic medication compared to those starting a beta-3 agonist. Methods We conducted a population-based, retrospective, matched cohort study using linked administrative data from Ontario, Canada from 2010 to 2018. We matched 47 324 new users of anticholinergic medications (oxybutynin, tolterodine, solifenacin, darifenacin, fesoterodine, trospium) to 23 662 new users of a beta-3 agonist medication (mirabegron); all of the included medications are only indicated for the treatment of OAB. We measured 75 baseline variables (including comorbid conditions, recent medications, and prior healthcare utilization) and used these to create a propensity score; groups were similar across all measured variables after matching. The primary exposure was the class of OAB medication (anticholinergic or beta-3 agonist). The primary outcome was dementia using a validated administrative data definition. Results The most common anticholinergics used were tolterodine (40%), oxybutynin (29%) and solifenacin (26%). The median (interquartile range [IQR]) prescription duration of anticholinergics was 30 (30-170) days. The median (IQR) prescription duration of a beta-3 agonist (mirabegron) was 64 (30-317) days. There was an increased risk of dementia among anticholinergic users compared to beta-3 agonist users (hazard ratio 1.23, 95% confidence interval 1.12-1.35). There was a significant effect modification based on both gender and age; men and those aged <= 75 years on anticholinergics had the highest risk of dementia relative to similar beta-3 agonist users. Conclusions The use of anticholinergic medications among patients with OAB was associated with an increased risk of new-onset dementia compared to beta-3 agonist users. These results address the potential protopathic bias present in other studies on this topic and support the association between anticholinergic medication use and dementia.