Drugs with anticholinergic activity are widely used to treat various conditions, including urinary incontinence and respiratory, neurological, and psychiatric disorders. Their mechanism of action is based on inhibition of acetylcholine receptors, which play a key role in learning, memory, and executive function. Chronic inhibition of their action has been linked to an increased risk of mild cognitive impairment and dementia.1,2 Numerous studies have suggested that cumulative anticholinergic burden may determine the progression of cognitive impairment,2,3 highlighting the need for careful assessment when prescribing these medicines, particularly in vulnerable patients such as older adults.1,4

Patients with neuropsychiatric disorders are characterised by high exposure to psychotropic anticholinergic drugs and frequent polypharmacy, even in middle age. In this context, accumulated anticholinergic burden can reach clinically relevant levels. Furthermore, in these patients, cognitive impairments may be part of the baseline clinical picture or coexist with it, making it difficult to identify potentially reversible pharmacological effects. Assessing anticholinergic exposure and burden in this population is of particular clinical interest, as it may help to improve the interpretation of cognitive symptoms, optimise prescribing, and prevent avoidable cognitive decline.

Various scales have been developed to classify anticholinergic burden, including the Anticholinergic Drug Scale,5 the Anticholinergic Risk Scale,6 and the Anticholinergic Cognitive Burden Scale (ACB),7 the latter being among the most widely used scales in clinical research.8,9 These tools estimate the total anticholinergic burden to which patients are exposed and can help clinicians make decisions that reduce its negative impact on cognition.10

The aim of this study was to determine the rate of anticholinergic prescribing in a sample of patients with various neuropsychiatric disorders, as well as the prevalence of cognitive impairment in those taking anticholinergic drugs or exposed to a high anticholinergic burden according to different classification scales.

This study had an observational cross-sectional analytical design. It included 135 patients aged over 18 years of both sexes, admitted to the Sant Joan de Déu Terres de Lleida psychiatric hospital, and was conducted between March 2022 and March 2024. The protocol was approved by the institutional ethics committee, and informed consent was obtained from all participants. Patients were selected consecutively. The inclusion criterion was admission for at least 3 months, thereby ensuring stability in their medication regimen and psychopathological status. Exclusion criteria were as follows: patients who did not meet the inclusion criterion, those for whom cognitive assessment could not be performed, and those for whom study data were missing.

Given the lack of a universally accepted reference scale for estimating anticholinergic burden and the incomplete inclusion of relevant drugs in some scales, we designed an exploratory composite index to maximise documentation of prescribed anticholinergic drugs in our sample. This approach addressed a descriptive and pragmatic need, aiming to more fully reflect actual pharmacological exposure in neuropsychiatric patients. The composite index was based on 14 scales for classifying anticholinergic burden (Table 1). Using all 14 scales, we compiled a list of 168 anticholinergic drugs, including their Anatomical Therapeutic Chemical Classification System code, pharmacological family, and anticholinergic burden. The mean of the scale scores was calculated where there were differences. This method has been used in the development of other scales that combine multiple scales, such as the CRIDECO Anticholinergic Load Scale.11 As this study had a descriptive design, no prior sample size calculation or assessment of statistical power was performed to evaluate the performance or validity of the composite index.

At the time of each patient's inclusion, we recorded all anticholinergic drugs active in the hospital pharmacy's electronic prescription system and that were included in the composite index (a total of 168 drugs), and conducted a baseline cognitive assessment using the Mini-Mental State Examination (MMSE). A score of 26 points or lower was used as a cut-off for cognitive impairment.22 In the final analysis, we calculated the anticholinergic burden using the ACB scale, which includes 88 drugs, and the 14-scale composite index, which incorporates the ACB scale.

All scales categorised the drugs using the following values: 0 = no anticholinergic activity; 1 = low anticholinergic burden; 2 = moderate; and 3 = high. A high anticholinergic burden was defined as 3 points or more.

The ACB scale was selected as the primary comparator because it is one of the most widely used and validated tools for estimating anticholinergic burden in relation to cognitive function. Its widespread use in clinical and epidemiological studies facilitates comparison with previous literature and supports its applicability in real-world clinical settings, including hospitals. However, the Drug Burden Index, which accounts for dose in estimating drug burden, was not included because detailed information on daily doses was not available in the electronic prescribing system.

For each patient, sociodemographic variables (sex, age, and educational level) and functional status were recorded using the Barthel Index and Reisberg's Global Deterioration Scale. Information was also collected on the history of substance use, main organic and psychiatric diagnoses according to the Diagnostic and Statistical Manual of Mental Disorders, 5th edition, the presence and aetiology of cognitive impairment, and the MMSE score at the time of study inclusion. Additionally, prescribed anticholinergic drugs were recorded, along with their individual and total anticholinergic burden. All assessments were conducted by the same researcher.

Cognitive impairment was evaluated using 2 complementary approaches. Firstly, we recorded whether the patients had a prior clinical diagnosis of cognitive impairment. This group represents a population that is particularly vulnerable to the adverse effects of anticholinergic burden. Secondly, we defined cognitive impairment based on MMSE score using a cut-off of 26 points or lower, thereby identifying patients with potentially undiagnosed cognitive impairment.

We conducted a descriptive analysis of the sample. Quantitative variables are expressed as measures of central tendency and dispersion, and qualitative variables are expressed as frequencies and percentages. The chi-squared test was used to assess the association between cognitive impairment (yes/no) and anticholinergic burden (<3/≥3). The odds ratio (OR) with a 95% confidence interval (95% CI) was calculated to estimate the effect. The t test was used to compare anticholinergic burden as a continuous variable between patients with and without cognitive impairment, and to assess the relationship between MMSE score and dichotomous anticholinergic burden (<3/≥3). Pearson's correlation was used to evaluate the association between MMSE score and continuous anticholinergic burden. Finally, a multiple linear regression model was constructed to assess the mean change in MMSE score per unit of anticholinergic burden as a continuous variable, adjusting for sex, age, educational level, substance use, and a history of hypertension, diabetes mellitus, and depression. A P-value of <0.05 was used as a cutoff for statistical significance. Data analysis was performed using PSPP version 2.0.0 software.

Table 2 shows the sociodemographic and functional characteristics of the study patients. Table 3 shows the clinical characteristics of the sample. Substance use was documented in 60 patients (44%). The most commonly used substance was alcohol (42.2%), followed by cocaine (17%) and cannabis (16.3%).

All patients received at least 1 anticholinergic drug, with a mean of 4.1 ± 1.7 drugs per patient. The maximum number of anticholinergic drugs was 10. The supplementary material lists the anticholinergic drugs used in our sample.

According to the ACB scale, the mean anticholinergic burden was 4.8 ± 2.5 (range 0–15). Cognitive impairment was observed in 76 of 106 patients in the high-burden group and in 15 of 29 in the low-to-moderate-burden group (Fig. 1). The difference was statistically significant (χ2 = 4.13; p = 0.042; OR = 2.36; 95% CI: 1.02–5.49).

Figure 1.

Percentage of patients with or without cognitive impairment according to anticholinergic burden on the ACB scale. *p < 0.05.

On the set of 14 scales, the mean anticholinergic burden was 5.9 ± 2.8 (range 1–19). Cognitive impairment was observed more frequently in the high-burden group (N = 120), as shown in Fig. 2, without reaching statistical significance (χ2 = 3.30; p = 0.069; OR = 2.36; 95% CI: 0.92–6.10).

Figure 2.

Percentage of patients with or without cognitive impairment according to low-to-moderate vs high anticholinergic burden on the set of 14 scales. AC, anticholinergic.

The ACB scale was used to compare anticholinergic burden between patients with cognitive impairment (mean = 5.16; SD = 2.53) and those without (mean = 4.09; SD = 2.29). The difference was statistically significant (t = −2.38; p = 0.019), with a lower mean burden in the patients without cognitive impairment (mean difference = −1.07; 95% CI: −1.97 to −0.18).

Using the composite index, the mean difference in anticholinergic burden between groups was −0.80 (95% CI: −1.81 to 0.21), without reaching statistical significance (t = −1.56; p = 0.121).

MMSE scores were compared between patients with low-moderate burden (mean = 26.14; SD = 3.74) and high burden (mean = 25.49; SD = 3.98), as defined by the ACB scale. No statistically significant difference was observed between groups (t = 0.79; p = 0.433). The mean difference between groups was 0.65 points (95% CI: −0.98 to 2.28).

Using a composite index, a statistically significant difference in MMSE scores was observed between the low-moderate group (mean = 27.20; SD = 2.27) and the high-burden group (mean = 25.43; SD = 4.05) (t = 2.55; p = 0.017). The mean difference between groups was 1.77 points (95% CI: 0.34–3.19), suggesting that patients with a low anticholinergic burden performed better cognitively.

Using the ACB scale, higher anticholinergic burden was significantly associated with lower MMSE scores (r = −0.178; p = 0.039). However, no statistically significant relationship was found between anticholinergic burden, calculated using the composite index, and cognitive performance on the MMSE (r = −0.092; p = 0.289).

Subsequently, multiple linear regression analysis was performed to assess the association between anticholinergic burden, calculated using the ACB scale, and MMSE score. In the model adjusted for sex and age, a significant negative association was found between anticholinergic burden and cognitive performance. However, after adjusting for educational level, history of substance use, and medical comorbidities, this association lost statistical significance. Table 4 shows the full results of the regression models.

Using the composite index, the multiple linear regression model adjusted for sex and age (Table 5) revealed no significant associations.

Among our patients with neuropsychiatric disorders, we observed a high rate of anticholinergic drug prescriptions, with a mean consumption of 4.1 ± 1.7 drugs per patient. By combining 14 different scales, we identified a larger number of anticholinergic drugs, resulting in a higher estimated total burden than that obtained using the ACB scale alone (mean 5.9 vs 4.8 points). Approximately 70% of the sample had a score of 3 or higher, indicating that most patients were exposed to potentially harmful levels of these drugs. Our findings are consistent with those reported in previous studies. It is common to find high use of anticholinergic drugs in patients with this profile, with approximately half of them having a high anticholinergic burden.23,24

Using the ACB scale to calculate burden, we observed a significant negative association with cognitive performance. In contrast, the composite index produced generally nonsignificant results. One factor that might explain the lack of a significant negative association when using the composite index is that this approach may produce inconsistency in the classification of anticholinergic burden. This is because the scales differ in terms of their criteria, level of evidence, and assessment of specific medications.25 Furthermore, the ACB scale was developed specifically to assess the impact of anticholinergic drugs on cognitive function,26 and may therefore be more specific in detecting associations in this domain than a broader composite index. In this regard, although the anticholinergic burden calculated using the 14-scale composite index better reflects total pharmacological exposure, it may not be the best predictor of cognitive decline if it includes medications with limited clinical relevance in this setting.27

In the multivariable models, anticholinergic burden, as measured on the ACB scale, was associated with a significant reduction in MMSE scores in the model adjusted for sex, age, and educational level, but this association lost significance after adding the remaining covariates. The composite index showed similar negative coefficients, but did not reach statistical significance in the various models.

Despite the methodological limitations of the composite index, it is noteworthy that the association between anticholinergic burden and cognitive performance followed the same direction as that observed with the ACB scale. The lack of statistical significance likely reflects the heterogeneity of the included scales, the presence of drugs with limited clinical relevance, and the statistical power constraints of the study. These results support the future development of composite indices based on robust pharmacological criteria and subjected to formal validation processes.

Multiple studies have demonstrated that anticholinergic drugs have clinically relevant effects on cognition, particularly with cumulative exposure or in vulnerable patients, especially those aged 65 and over.1,2 In particular, when measured using the ACB scale, higher anticholinergic burden has been associated with lower MMSE scores26,28; patients with high anticholinergic burden have shown reductions of between 2.526 and 6.4 points29 on the MMSE. Although these effects are substantially greater than those observed in our study, they support the finding that high anticholinergic exposure is associated with reduced cognitive function. The observed differences may be because the participants in these studies were all over 65 years of age, with a mean age of over 80. By contrast, our participants were aged between 28 and 86 years, with a mean age of 59 years.

In this context, to mitigate the potential adverse effects of anticholinergic drugs on cognitive function, we emphasise the need to review prescriptions and, whenever possible, to limit the use of drugs with a high anticholinergic burden in patients with cognitive impairment or at risk of developing it. Optimising pharmacological treatment in this patient group could help improve clinical outcomes and minimise the risk of further cognitive decline. Specific strategies include conducting regular reviews of prescriptions, identifying drugs with a high anticholinergic burden that could be deprescribed or replaced with safer alternatives, and conducting an individualised assessment of the risk–benefit ratio for each medication.

This study is limited by its observational, cross-sectional design, which prevents the establishment of causal relationships. We cannot determine whether exposure to anticholinergic drugs preceded cognitive decline or whether there is indication bias. The absence of a predefined sample size limits our ability to state with certainty that the lack of statistical significance observed in some comparisons reflects an actual absence of association rather than insufficient power to detect small or moderate effects. Given that formal validation of a new index requires a study design and sample size sufficient to ensure adequate statistical power, we suggest that future studies aimed at validating the composite index use prospective sampling with an a priori power calculation and adjust for key covariates. Therefore, we present the composite index as an exploratory tool for describing pharmacological coverage, and not as a validated scale. It should also be noted that, although the MMSE is a practical and widely used tool, it has limited sensitivity to mild cognitive changes and may be influenced by educational level and other sociodemographic factors.30

Our findings highlight the importance of considering not only how comprehensively anticholinergic burden is measured, but also the specificity of the scales used, particularly when analysing its relationship with cognitive performance. Future studies could evaluate whether a combination of scales, weighted appropriately, optimises the detection of effects on cognition. Standardised and validated tools for assessing anticholinergic burden that cover a wide range of relevant drugs while maintaining specificity would be highly useful in improving their applicability in clinical practice. Furthermore, future prospective trials or longitudinal studies are needed to clarify the potential causal relationship between anticholinergic burden and cognitive decline.

In conclusion, our results support the hypothesis that exposure to anticholinergic drugs is associated with reduced cognitive performance. Furthermore, they suggest that greater pharmacological coverage of the scales does not necessarily improve their predictive power with regard to cognition if clinical specificity is compromised. The documented high rate of cumulative exposure to anticholinergic drugs among hospital and neuropsychiatric patients underscores the need for ongoing education of healthcare professionals about the risks of prescribing these drugs, the regular review of pharmacological guidelines, and the implementation of deprescribing strategies or substitution with safer alternatives whenever feasible.

This study compares 2 measures of anticholinergic burden: the Anticholinergic Cognitive Burden Scale, which includes 88 drugs, and a composite index of 14 scales, which includes 168 drugs. Although the composite index provides a more comprehensive assessment of anticholinergic exposure, the results show that the burden measured by the index was not statistically significantly associated with cognition, highlighting the importance of scale specificity over mere coverage.

The results support the thorough review of anticholinergic drugs and the prioritisation of deprescribing strategies in patients with polypharmacy. Furthermore, they encourage the research, validation, and standardisation of tools that can comprehensively document anticholinergic drugs with specific effects on cognition in order to optimise clinical practice.

The study was approved by the Clinical Research Ethics Committee of the Sant Joan de Déu Terres de Lleida psychiatric hospital (code 04–2022-03) on 17 March, 2022. All participants signed the informed consent form.

None declared.

María Ruiz-Julián: Writing – original draft, Software, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualisation. Gerard Piñol-Ripoll: Writing – review & editing, Validation, Supervision, Conceptualisation.

None declared.