Neurogenic lower urinary tract dysfunction (NLUTD) is a common complication in individuals with conditions such as spinal cord injuries, multiple sclerosis (MS) and other neurologic disorders, often presenting with symptoms of overactive bladder, incontinence and voiding dysfunction (1,2). Anticholinergic medications, which inhibit muscarinic receptors to reduce detrusor muscle overactivity, are frequently employed as first-line pharmacological interventions to manage these symptoms (1,3). However, the widespread distribution of muscarinic receptors in the central nervous system raises concerns regarding potential cognitive side effects, particularly in vulnerable populations (2,4).

Cognitive impairments, including memory loss, attention deficits and decreased processing speed, have been reported with anticholinergic use, especially in older populations (2,5,6). Studies show that anticholinergic medications cross the blood-brain barrier, contributing to brain atrophy and increased dementia risk (2,7). A cohort study by Wang et al (2) found that cumulative exposure to bladder-specific anticholinergics was associated with an elevated risk of dementia, emphasizing the need for careful therapeutic balancing between symptom management and cognitive safety.

In cognitive research, functions are commonly categorized into key domains: Memory, attention and processing speed, executive functions, visuospatial ability and global cognition (8). NLUTD studies often use standardized tools to assess these domains. For instance, the Mini-Mental State Examination (MMSE) is a 30-point scale examining orientation, recall, attention, language and constructional ability (9). The Symbol Digit Modalities Test (SDMT) is widely used, particularly in neurological populations, to evaluate attention and processing speed (10). More specialized tests, such as the Frontal Assessment Battery (FAB) or the Trail Making Test, are used to evaluate executive functioning, and instruments such as the Montreal Cognitive Assessment (MoCA) assess multiple domains, including visuospatial abilities (9). By briefly outlining these tools and domains, a clear conceptual framework that supports analyses of cognitive outcomes was established.

In populations with neurogenic bladder dysfunction, the cognitive burden of anticholinergics may be particularly problematic due to pre-existing neurological vulnerabilities. Sakakibara et al (4) demonstrated that although some anticholinergic agents, such as imidafenacin, may present lower cognitive risks, others such as oxybutynin and tolterodine have been linked to significant declines in memory and executive function. Similarly, Morrow et al (7) highlighted a substantial impact on processing speed and memory in patients with MS treated with anticholinergic medications, with impairments detected using standardized cognitive tests.

Despite their therapeutic efficacy in alleviating NLUTD symptoms, the long-term cognitive consequences of anticholinergic medications remain a critical area of investigation. Emerging evidence suggests that alternative treatments, such as β3-adrenergic agonists like mirabegron, may provide effective symptom control with fewer cognitive side effects, as demonstrated in comparative studies (1). Trbovich et al (1) found improved cognitive outcomes and comparable efficacy when switching from anticholinergics to mirabegron in individuals with neurogenic bladder dysfunction.

The present systematic review and meta-analysis aim to comprehensively evaluate the cognitive effects of anticholinergic medications in patients with NLUTD, synthesizing data from clinical trials and observational studies. By examining the extent and nature of cognitive impairments across various populations and drug types, the present study seeks to inform clinical decision-making and highlight potential therapeutic alternatives that minimize cognitive risks.

The present systematic review and meta-analysis investigated the cognitive effects of anticholinergic medications in individuals with NLUTD. The primary findings suggest a trend toward negative cognitive impact-particularly in the domains of memory, executive function and attention-though the overall effect using a random-effects model did not reach statistical significance. This discussion contextualizes these findings within the broader literature, highlights potential explanations for the heterogeneous outcomes, and proposes clinical implications and future directions.

A key observation from the pooled results is the modest, albeit statistically non-significant, negative overall impact of anticholinergics on cognitive performance (SMD=-0.33; 95% CI: -0.72 to 0.06; P=0.10). Although the fixed-effects model produced a statistically significant SMD of -0.178 (P=0.011), the random-effects model was prioritized, which yielded a non-significant SMD of -0.33 (P=0.10). This choice was made because the random-effects model incorporates between-study heterogeneity, combining both within-study and between-study variance into the pooled estimate-an approach that is most appropriate given the observed high heterogeneity (Q=162.25, P<0.001; τ²=0.738) (15). By contrast, the fixed-effects model assumes all studies estimate the same true effect. While it offers greater statistical power, it may underestimate uncertainty when substantial between-study variability exists (15). Therefore, to provide a more conservative and realistic estimate that reflects clinical diversity, the random-effects model was deemed more appropriate.

Our findings align with existing evidence indicating that anticholinergic medications can impair central cholinergic pathways involved in cognition (1,2). In particular, the cholinergic deficit hypothesis has been well-documented in several neurocognitive disorders, including Alzheimer's disease, emphasizing the importance of central acetylcholine in learning, memory, and attention (16). Given that anticholinergics block muscarinic receptors, a reduction in central cholinergic signaling can manifest as deficits in processing speed, executive functioning, and memory formation (17). While the lack of significance in the present meta-analysis calls for caution in interpreting the data, the consistency with theoretical mechanisms and other clinical trials warrants continued scrutiny of anticholinergic-induced cognitive decline.

A striking feature of the present meta-analysis is the substantial heterogeneity (Q-value of 162.25, P<0.001; τ²=0.738). The wide prediction interval (-2.17 to 1.51) highlights significant between-study variability and reflects instability in the overall effect. This contrasts with the narrower CI, which estimates the precision of the mean effect (18). The breadth of the interval indicates that findings should be interpreted cautiously (19), and highlights the need for future large-scale and standardized trials to reduce uncertainty and improve estimation of the true cognitive impact of anticholinergics. Several factors might underlie this variability in effect sizes. First, differences in sample characteristics, such as underlying etiologies (for example, MS, SCI and OAB) and disease progression, could contribute to varying baseline vulnerability to cognitive dysfunction. Individuals with MS or SCI often exhibit pre-existing cognitive deficits due to demyelination or traumatic central nervous system injury (20,21). The incremental burden of anticholinergic medications may, therefore, differ depending on neurological reserve and the severity of baseline impairments. Second, the use of various anticholinergic agents (for example, oxybutynin, tolterodine, solifenacin and imidafenacin) with diverse pharmacokinetic profiles likely influenced cognitive outcomes. Older-generation medications such as oxybutynin and tolterodine are known to cross the blood-brain barrier more readily, thereby increasing the likelihood of central side effects (22). Conversely, newer or more selective agents (for example, solifenacin and imidafenacin) may exert a lesser central anticholinergic burden, resulting in lower cognitive risk (22). It is thus plausible that differences in agent selectivity, dosing schedules, and treatment durations contributed to the heterogeneity detected in this review. Third, variations in outcome measures across the included studies can also drive heterogeneity. Some studies relied on global cognitive measures (for example, MMSE, ADAS-cog, FAB), which may lack sensitivity to subtle cognitive changes. More comprehensive or domain-specific instruments (for example, the SDMT or detailed memory tasks) might detect more nuanced impairments. This meta-analysis found stronger associations in targeted domains such as memory, executive function, and attention, suggesting that future research should prioritize sensitive, domain-specific assessments to capture the full extent of cognitive changes induced by anticholinergics.

Despite the non-significant pooled effect, the subgroup analyses offered valuable insights. Attention, executive function, and memory domains emerged as particularly vulnerable to the effects of anticholinergics. These domains are heavily reliant on adequate cholinergic transmission (23). Studies utilizing tests such as the Stroop Test, the SDMT, and word-list learning tasks have repeatedly demonstrated that blocking cholinergic receptors impairs selective attention, mental flexibility, and the encoding of new memories (24,25). In NLUTD populations where patients may already contend with the cognitive burden from neurological disorders, even minor reductions in performance can have profound practical consequences, affecting daily living activities, adherence to rehabilitation protocols, and overall quality of life.

The moderator analyses underscore the complexity of factors influencing cognitive vulnerability. Intriguingly, a higher proportion of females in a study's cohort correlated with smaller negative cognitive effects (P<0.001). Although evidence is mixed, some literature suggests that sex-related hormonal and genetic factors could modulate pharmacodynamic and pharmacokinetic responses to medications (26). Similarly, older age was associated with reduced severity of cognitive decline, suggesting that younger individuals with NLUTD may be more susceptible (P<0.001). While this finding appears counterintuitive-given that older adults are generally considered at elevated risk of anticholinergic-induced delirium or cognitive impairment-certain explanations may apply. Younger cohorts might exhibit higher baseline functioning, making any decrement from baseline more pronounced. Additionally, older adults with significant frailty or comorbidities might be preferentially prescribed lower doses or more selective anticholinergics, thereby moderating their cognitive risks.

Clinical implications from these findings revolve around two major points. First, prescribers must balance therapeutic goals in managing NLUTD with the potential for adverse cognitive events. Urinary incontinence, frequency, and urgency negatively affect patient well-being, but the costs of therapy include potential detrimental effects on higher-order brain function. In populations with existing neurologic compromise, vigilance is paramount. Clinicians could consider minimizing exposure to highly lipophilic or non-selective anticholinergics and employing the lowest efficacious dose. Second, alternative therapies such as β3-adrenergic agonists (for example, mirabegron) have shown promise in providing symptomatic relief with less central cholinergic interference (27). Transitioning or combining these agents with other interventions-for example, pelvic floor rehabilitation and neuromodulation therapies-may be a viable strategy for individuals with significant cognitive vulnerability (28).

While the publication bias analysis and trim-and-fill adjustments indicate only a minor impact of potential unpublished negative or null studies, the slight asymmetry in the funnel plot underscores the need for continued, transparent reporting of cognitive outcomes in NLUTD populations. Researchers must publish all relevant data, including neutral or equivocal results, to ensure that meta-analytic estimates accurately reflect true effects. Although Begg's test indicated possible small-study effects, the non-significant Egger's result and minimal change after applying trim-and-fill (only two studies imputed, with little effect on the pooled SMD) suggest that publication bias likely does not materially alter our overall conclusions (29).

The present meta-analytic systematic review has several limitations. First, the overall number of included studies was small (N=3), limiting the power to detect modest but clinically relevant cognitive changes. Moreover, the short treatment durations and heterogeneous follow-up periods (ranging from a few weeks to months) may not fully capture the long-term cognitive trajectory of patients receiving anticholinergics. Second, the included studies employed various cognitive tests, ranging from global measures (for example, MMSE) to domain-specific tools, making direct comparisons challenging and introducing potential measurement bias. Third, while the meta-analysis attempted to evaluate publication bias, the limited number of studies constrains the reliability of these statistical tests. Fourth, confounding variables-such as concomitant medication use, overall anticholinergic burden from multiple sources, and comorbid psychiatric or neurological conditions-were not consistently reported or controlled for across studies. Fifth, most studies had relatively small sample sizes and did not stratify participants based on the severity of NLUTD or baseline cognitive function, precluding more nuanced subgroup analyses. Sixth, although the subgroup analyses revealed statistically significant SMDs for attention (0.752), executive function (-0.948), and memory (-0.645), it was emphasized that these results are preliminary. With only three studies and small subgroup samples, estimates may be unstable, and type I error is possible (30). These findings should be considered hypothesis-generating rather than confirmatory, aligned with broader trends observed in anticholinergic research that highlight domain-specific vulnerabilities but await validation in larger, adequately powered studies. Seventh, the moderator findings-suggesting less cognitive impact in studies with more female participants and greater declines in younger cohorts-should be interpreted cautiously. With only three contributing studies, the estimates are statistically underpowered and potentially confounded, in line with methodological cautions against overinterpreting such results. As such, these results remain hypothesis-generating and should be confirmed in larger, more robustly powered analyses. Eights, the subgroup and moderator analyses were performed post hoc and not pre-specified in a formal protocol. As such, these findings should be interpreted with caution, recognizing their exploratory nature. Finally, the observed high heterogeneity (Q=162.25, τ²=0.738) likely reflects not only clinical diversity but also methodological variability-notably, differences in cognitive testing instruments and anticholinergic pharmacodynamics. For instance, more sensitive domain-specific tools (for example, SDMT, FAB) may detect subtle deficits missed by global measures such as MMSE. Additionally, drug-specific factors-such as muscarinic receptor subtype selectivity and central nervous system penetration-might explain variability, as older agents (for example, oxybutynin) pose higher cognitive risks than newer, more selective agents (for example, imidafenacin) (31,32). Future meta-regression or individual participant data analyses are needed to more precisely quantify these structural sources of heterogeneity.

In conclusion, the present meta-analysis suggests that anticholinergic medications may exert a negative, though variably expressed, impact on cognition in patients with NLUTD, particularly in memory, executive function, and attention. While the pooled effect did not reach statistical significance using a random-effects model, considerable heterogeneity points to the influence of patient characteristics, drug choice, and methodological differences in outcome assessment. These findings underscore the need for carefully weighing therapeutic benefits against potential cognitive risks, especially in populations already burdened by neurological compromise. Future large-scale, longitudinal studies with standardized cognitive measures, detailed patient stratification, and attention to confounding factors are essential to refine our understanding of anticholinergic-related cognitive changes and guide safer prescribing practices.