
β3 Agonists Show No Dementia Advantage Over Anticholinergics in OAB
Key Takeaways
- A national EHR cohort compared first-line OAB drug classes: high-risk anticholinergics (68.9%), low-risk anticholinergics (6.0%), and β3 agonists (25.1%), with 7.8% developing dementia.
- Weighted models showed high-risk anticholinergics had lower dementia hazard versus β3 agonists (HR 0.90), while low-risk anticholinergics were not significantly different (HR 0.97).
β3 agonists were not linked to lower dementia incidence vs anticholinergics among 830,393 adults treated for overactive bladder, a cohort study found.
Adults treated for
Study Compared 3 Treatment Classes
Concern that anticholinergics raise dementia risk has pushed clinicians toward β3 agonists, which are presumed to be cognitively safer. To test that assumption, researchers at University Hospitals Cleveland Medical Center and Oberlin College used Epic Cosmos electronic health record (EHR) data from 2012 to 2025. The cohort included 830,393 adults aged 40 to 90 years who started drug therapy within 1 year of an OAB diagnosis and had no prior dementia. The median age was 67 years, and 70.5% were women.
Patients were grouped by the first medication class prescribed. A total of 572,489 (68.9%) started high-risk anticholinergics (oxybutynin, tolterodine, or solifenacin); 49,592 (6.0%) started low-risk anticholinergics (darifenacin, trospium, or fesoterodine); and 208,312 (25.1%) started β3 agonists (mirabegron or vibegron). Dementia, identified by diagnosis codes or prescriptions for dementia medications, developed in 65,134 patients (7.8%). Median follow-up ranged from 1046 days for β3 agonists to 1568 days for high-risk anticholinergics.
High-Risk Anticholinergics Showed a Slightly Lower Risk
After weighting to balance patient characteristics, high-risk anticholinergics were associated with a lower HR of dementia than β3 agonists (HR, 0.90; 95% CI, 0.87-0.93). Low-risk anticholinergics did not differ significantly from β3 agonists (HR, 0.97; 95% CI, 0.93-1.01). Weighted 10-year dementia risk was 20.0% with β3 agonists, 19.0% with low-risk anticholinergics, and 18.1% with high-risk anticholinergics.
In the discussion, the authors wrote, "In this large, national, EHR-based cohort study of over 830,000 adults with OAB, we found no evidence that β3 agonists were safer compared with anticholinergics with respect to incident dementia."
The investigators said the small gap was most consistent with residual channeling bias, in which clinicians following guideline advice may steer older adults or those at higher cognitive risk toward β3 agonists. E-values, which estimate how strong unmeasured confounding would need to be to nullify a result, were low at 1.29 (high-risk anticholinergics vs β3 agonists) and 1.20 (high-risk vs low-risk anticholinergics). Modest confounding could therefore shift the 95% CIs to include no difference.
Subgroup results were inconsistent. Among women younger than 65 years, low-risk anticholinergics carried a higher risk than β3 agonists (HR, 1.20; 95% CI, 1.03-1.39), but no subgroup showed a pattern consistent with anticholinergic neurotoxicity.
Duration Not Linked to Higher Risk Through 400 Days
An exploratory landmark analysis of 395,709 patients found no evidence that longer exposure, up to 400 days, increased dementia incidence. The approach starts follow-up after an initial treatment period to limit immortal time bias, which can make longer treatment look protective. For high-risk anticholinergics, each additional 100 days of use was associated with a slightly lower dementia rate (log HR slope per 100 days, −0.022; 95% CI, −0.045 to −0.001). Slopes for low-risk anticholinergics and β3 agonists did not differ from zero.
The results diverge from a population-based Ontario study that matched 47,324 new anticholinergic users with 23,662 new mirabegron users and reported a higher dementia risk with anticholinergics (HR, 1.23; 95% CI, 1.12-1.35).2 The Ontario authors said their design addressed protopathic bias. The Cleveland team said differences in study design, including how immortal time bias and exposure were handled, likely explain some of the divergence from earlier work.1 They also proposed that OAB may itself be an early sign of the neurodegenerative and cerebrovascular processes that lead to dementia.
Limitations Temper Conclusions
Exposure was based on prescription orders rather than dispensing, dose data were often unavailable, and dementia events outside the Cosmos network were not captured. The dementia definition may also underestimate mild or early-stage disease. The authors noted that the retrospective design precludes a definitive conclusion about medication class and dementia risk, although they concluded that any class-specific contribution to dementia risk was small.
References
1. Sheyn D, Rhodes S, Mohamed Z, Hijaz A. Drug therapies and incidence of dementia among adults with overactive bladder. JAMA Netw Open. 2026;9(11):e2642026. doi:10.1001/jamanetworkopen.2026.42026
2. Welk B, McArthur E. 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. BJU Int. 2020;126(1):183-190. doi:10.1111/bju.15040
Related to this article








