Anticholinergic medications—which exert their effects by blocking the action of the neurotransmitter acetylcholine—are prescribed to treat a wide range of medical conditions. An accumulating base of literature suggests that chronic treatment with these agents could lead to an increased risk of cognitive impairment or dementia, prompting some groups, such as the American Geriatrics Society, to recommend restricting their use. Further studies have shown that the risk posed by use of these agents is dependent on cumulative exposure over time, with longer durations of treatment presenting a greater risk of negative effects on cognition. Antimuscarinic drugs, used to treat overactive bladder specifically, have very high anticholinergic activity, and many have precautions included in their prescribing information regarding central nervous system anticholinergic effects. Given the potential for symptom exacerbation, the oxybutynin label additionally recommends caution for use in patients with preexisting dementia treated with cholinesterase inhibitors. 

Results of a recent meta-analysis of short-term anticholinergic use (average follow-up duration of 12 weeks across studies) by Salahia and colleagues [1] suggest that perhaps the link between anticholinergics and cognitive impairment/dementia has not been fully established, although their review was limited by only one study available to assess the outcome of dementia. The current systematic literature review and meta-analysis assessed the impact of ≥3 months of anticholinergic use on the risk of incident dementia (all subtypes), incident mild cognitive impairment, and change in cognitive function across all patients, as well as specifically in patients treated with anticholinergic medications for overactive bladder.

This systematic literature review and meta-analysis utilized the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement for systematic reviews and meta-analyses, and the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) reporting guidance. The protocol was registered in the PROSPERO database. PubMed, Embase, and Cochrane Library databases were used to identify studies published prior to August 8, 2019. All articles were independently screened and assessed for quality by two reviewers. Data abstraction was conducted by one reviewer using a standardized data abstraction form, and data were verified by the second reviewer against the original publications.

Titles and abstracts were reviewed and moved to the full-text review if they met the following inclusion criteria: (1) examined the impact of ≥3 months of anticholinergic drug use on dementia or cognitive function in adult patients (18+ years of age), (2) was a randomized-controlled trial, case-control study, or cohort study, (3) contained an adequate description of the methods used, and (4) was a primary publication. Quality was assessed using the Cochrane Risk-of-Bias Tool for Randomized Trials (RoB 2) and the Risk Of Bias In Non-randomized Studies – of Interventions (ROBINS-I) for observational studies. 

All studies included in the qualitative review were considered for the meta-analysis if they met all other criteria and contained a comparison group with no anticholinergic use. Meta-analyses were conducted using random effects models. The 95% confidence intervals (CIs) and 95% prediction intervals (PIs) were reported. The CI reflects random error in estimating the mean but does not reflect the spread of the random-effects distribution. The PI reflects heterogeneity and random estimation error and may be informally interpreted as the interval within which we expect the true value estimated from a future study to lie. Influence analysis was performed to assess if any individual study had a particular impact on the results. The impact of study characteristics was explored using stratified and meta-regression analyses. Funnel plot asymmetry was also assessed. Analyses were conducted with STATA v.16 statistical software (Stata Corporation LP, College Station, TX).

A total of 2092 articles were retrieved based on the electronic search, with another 30 identified from a supplemental manual search. After deleting 132 duplicates, 1990 records were screened based on their titles and abstracts, and 316 were included in the full-text review. Finally, 21 studies were included for qualitative synthesis, with the majority reporting an increased risk with anticholinergic use overall or for ≥1 dosing exposure level (incident dementia: 8 of 9 studies; incident Alzheimer’s disease: 4 of 4 studies; incident mild cognitive impairment: 2 of 2 studies; cognitive impairment/decreased performance: 7 of 11 studies).

Six of 9 studies assessing incident dementia were considered relevant for inclusion in a meta-analysis. The study characteristics are summarized in Figure 1. Individual dementia subtypes, incident mild cognitive impairment and change in cognitive function could not be examined via meta-analysis given the limited number of articles.  

Across the 6 studies, the rate ratios (RR) for incident dementia ranged from 1.05 to 2.63, with an estimated average RR of 1.46 (95% CI, 1.17 to 1.81; 95% PI, 0.70 to 3.04; Figure 2). No single study was exceedingly influential on the results. When data from studies that assessed daily dosing information were examined, all dosing exposure levels were associated with increased rates of incident dementia, with higher dosing categories presenting greater risk (90-365 total standardized daily doses [TSDD]/defined daily doses [DDD] vs 0: 1.14 [1.07 to 1.21; n=3]; 365-1095 vs 0 TSDD/DDD: 1.35 [1.24 to 1.48; n=2]; >1095 vs 0 TSDD/DDD: 1.49 [1.38 to 1.61; n=2]). 

Subgroup analysis and meta-regression of study characteristics showed that the RRs were generally consistent with the main analysis. On average, positive associations were reported in case-control studies, studies with no lag, studies with a minimum age at enrollment of at least 65 years, studies with enrollment starting after 2000, studies with <70% female patients, and studies with ≥70% female patients. Begg and Egger tests yielded p values greater than 0.8 for asymmetry, and using the trim and fill method, the random-effects summary RR also was not dramatically affected (RR=1.63). Overactive bladder medications could not be examined in a separate meta-analysis given limited studies (n=2), but these nested case-control studies both reported an increased risk of dementia with ≥3-month use of bladder antimuscarinics (adjusted odds ratio range: 1.35 to 1.65 in Coupland [2] and 1.21 to 1.35 in Richardson [3]). Higher exposure categories appeared to present a greater risk (i.e., comparisons including >365 TSDD/DDD).

Use of anticholinergic agents for 3 months or longer was found to increase the risk of dementia an average of 46% relative to non-use. This increased risk also was reported in the two studies from the meta-analysis that evaluated anticholinergic medications used to treat overactive bladder. Based on subgroup analyses of study characteristics, the cognitive impact of anticholinergics appears to span age and sex groups, but additional studies are needed. Cumulative exposure modified the increased risk presented by anticholinergics in the current analysis, with greater increases at higher exposure categories.

Given the substantially increased risk of developing dementia associated with use of anticholinergic agents for ≥3 months, physicians should carefully weigh the risk versus the potential benefits prior to prescribing.

Salahia S et al. Anticholinergic drugs and risk of cognitive impairment or dementia in patients with overactive bladder syndrome: a systematic review and meta-analysis. Poster presentation at the International Continence Society 2019 Meeting; September 3-9, 2019; Gothenburg, Sweden.
Coupland CA et al. JAMA Intern Med. 2019;179(8):1084-1093.
Richardson K et al. BMJ (Clinical research ed). 2018;361:k1315.