People with psychosis (such as schizophrenia or bipolar with psychotic symptoms) have a significant burden of symptoms. Research has mainly focussed on helping symptoms such as voices and delusions (called ‘positive’ symptoms), and this is what most antipsychotic medication aims to improve.
However, we know from research that people with psychosis also have negative symptoms (low mood, lack of motivation and enjoyment) and cognitive impairment. This means they can have problems with memory, learning, information processing and problem-solving, which can massively impact their life. It is part of the reason that many people with chronic psychosis need community support, which represents a high proportion of healthcare costs.
The brain chemical acetylcholine helps with learning and cognitive processes and has become an interesting target for research. In Alzheimer’s dementia, for example, we increase the amount of acetylcholine with medication. However, some antipsychotics (such as clozapine and quetiapine) block receptors for acetylcholine and we commonly co-administer anti-cholinergic medication to reduce some of the side effects of our antipsychotic medications (such as stiffness or movement problems). This has the effect of reducing the amount of acetylcholine.
Taken together, one has to wonder whether some of the cognitive problems seen in psychosis are, in part at least, a consequence of the very medication we prescribe to help.
Mancini et al (2025) have conducted a study, recently published in the American Journal of Psychiatry, to see whether there may be an association between cognitive impairment and anticholinergic medication in psychosis.
People with experiences of psychosis can have problems with memory, learning, information processing and problem solving
Methods
The researchers undertook a systematic review and meta-analysis of relevant studies; this is a way of pooling results from numerous smaller studies. They searched three databases and used wide reaching search terms, including observational studies and clinical trials. They included studies involving adults over the age of 18 and diagnoses across the psychosis spectrum, including psychotic mood disorders and individuals at ultra-high risk of psychosis. The researchers excluded studies of patients with a primary disorder other than psychosis, such as anxiety, neurodevelopmental disorder or non-psychotic mood disorder. Studies were only included if they had objective quantitative measures of cognition, and so excluded studies that used self reporting measurements.
Potential studies were screened by independent investigators for suitability and analysed with random-effects meta-analysis modelling. This is a statistical method that takes into account heterogeneity, or variation, between the studies due to differences in methodology.
Results
40 studies were included in the meta-analysis. These fell into three categories:
- Studies assessing cognitive function and anticholinergic burden using clinical scales,
- Studies assessing cognitive function and anticholinergic burden using serological (lab based) measures and
- Studies comparing cognitive function before and after tapering or challenge with anticholinergics.
5,188 individuals were included in the analysis, of which most (87.6%) had a diagnosis of schizophrenia whilst the remainder had a diagnosis of first-episode psychosis. 64.5% of individuals were male, with an average age of 38.6.
The following statistically significant results were reported (with key results presented in greater detail):
- A negative correlation between anticholinergic burden and cognition
- Domains: global cognition (r=-0.37, 95% CI -0.48 to -0.25, pcorr
- 25 studies, n=4,620
- A negative correlation between serum anticholinergic activity and cognition
- Domains: verbal learning (r=-0.26, 95% CI -0.38 to -0.14, pcorr
- 6 studies, n=382
- Tapering off anticholinergic medications improved cognition
- Domains: verbal learning, working memory (d=0.94, 95% CI 0.63 to 1.26, pcorr = 0.001), and executive functions
- 9 studies, n=186.
A negative correlation means here that an increase in both anticholinergic burden and serum anticholinergic activity was associated with reduced cognition in the specific areas highlighted above.
As anticholinergic burden increased, global cognition reduced.
Conclusions
The authors summarise by saying:
In this systematic review and meta-analysis, we found that anticholinergic burden is associated with worse cognitive functioning in patients with psychosis and that reducing anticholinergic medication was effective in improving cognitive function. These findings highlight the negative impact of routinely used pharmacological interventions on cognitive function in individuals with psychosis. From a clinical perspective, tapering off anticholinergic medication may be beneficial. However, further randomised clinical trials are needed for an unbiased quantification of benefit.
Strengths and limitations
Overall, this study was well conducted and had high validity:
- The study question was well defined and used appropriate search terms (including clinical high risk of psychosis as part of the spectrum and psychotic mood disorders).
- Study identification was conducted by independent reviewers, and a method of adjudication was described for the case of disagreements with a clear PRISMA flow diagram, all of which enhanced internal validity, or the degree to which the study question was answered by the study.
- Most of the studies included were of high quality, with no publication bias found.
A key strength of this paper is that the authors conducted four sensitivity analyses and a meta-regression, which addressed the potential confounding issues of disease chronicity or antipsychotic dosage:
- The negative correlation was robust to all sensitivity analysis, including first-episode psychosis – meaning patients with long-term psychosis, or a single first episode, both had the same association between cognitive impairment and anticholinergics.
- Meta-regression found no statistically significant association between effect sizes and age, sex or chlorpromazine-equivalent antipsychotic dose, suggesting these factors do not explain the association.
One of the main limitations of this study is that it was unable to include other variables such as negative symptoms, depression or severity of psychotic symptoms, all of which may play a significant role in cognitive impairment. The authors point out the importance of future studies investigating and reporting on these potential confounders.
Additionally, the studies that involved tapering of anticholinergics did not include controls, so randomised controlled studies are needed to strengthen this finding.
Finally, despite referencing a wide range of psychotic disorders in their search terms, the papers included in this analysis only included diagnoses of schizophrenia and first episode-psychosis, so the results cannot necessarily be generalised to bipolar disorder with psychotic symptoms, or individuals at ultra-high risk of psychosis.
Overall, this study was well conducted and had high validity.
Implications for practice
This study has significant clinical implications for a potentially large number of patients. For decades research in schizophrenia and psychosis has focussed on targeting and reducing ‘positive’ symptoms, but slowly focus has shifted to lesser appreciated but significantly disabling domains of these illnesses – negative and cognitive symptoms. Cognitive symptoms can have wide reaching effects on people’s lives, contributing to high levels of unemployment and need for support in the community.
This paper was really exciting to read because its applications are so immediate and clinically relevant to day to day practice. Translational studies can seem abstract and removed from clinical work, when I know the benefits won’t be seen for decades but I have a patient I am concerned about in front of me. The patients I treat in south London are diverse and complex and there is a significant amount of socio-economic stress and deprivation that complicates any mental health presentation. This study presents a simple and safe way to potentially improve someone’s life. As the intervention is focused on reducing medication burden, it is likely to be popular with patients and carers, although it would be prudent to explore patients views on this.
All prescribing decisions should be made after careful risk vs benefits considerations, in joint discussions with patients. We don’t prescribe antipsychotics unless they are indicated – in my experience they are often transformative and life-saving. Additionally, anticholinergic medications can be vital in reducing uncomfortable and unpleasant side effects. However, this paper prompts us to question the burden of these medications and crucially when they should be reviewed.
Reviewing anticholinergic medication
Chengappa et al have written an editorial (Chengappa et al., 2025) about Mancini’s paper, that provides helpful suggestions about how a psychiatrist might go about reviewing anticholinergic medication burden (Chengappa et al., 2024). They suggest:
- Identifying patients who have been stable for at least six months and,
- with no extra-pyramidal side effects and,
- adopting a shared-decision making approach.
Reduction should be gradual (i.e. over months), with regular reviews to monitor psychotic symptoms, re-emergence of side effects and changes to cognitive functioning. It is a pragmatic and achievable approach, some of which is practised routinely by old age psychiatrists who are accustomed to measuring anticholinergic burden and reducing superfluous medication, and one that I believe can be embraced by general adult colleagues and the wider field of psychiatry.
Whilst this is all very encouraging, it is important to consider important caveats. This paper demonstrates correlation and further prospective controlled studies are required in order to demonstrate causation. Other elements that may impact cognition need to be considered in future studies, including the presence of negative symptoms and psychotic symptom severity.
This will not be a panacea, and for some people it will not be possible to reduce prescriptions, but emerging therapies such as KarXT may offer an alternative means to improve cognition.
Ultimately, I feel hopeful that cognitive impairment in chronic psychosis is getting the focus and action that it justifies and that patients will see real benefit in the not too distant future.
This paper prompts us to question the burden of these medications and crucially when they should be reviewed.
Statement of interests
Dr Dawkins has previously worked in the same research group as T Pillinger and R McCutcheon but had no involvement in the paper presented.
Links
Primary paper
Mancini V, Latreche C, Fanshawe JB, et al. (2025) Anticholinergic burden and cognitive function in psychosis: a systematic review and meta-analysis. Am J Psychiatry ; 182:349–359.
Other references
Chengappa K.N.R., Cotes R.O. (2024) Clozapine’s high incidence of ileus and pneumonia demand better clinical strategies—how do we get there? Am J Psychiatry, 181:851–853. 10.1176/appi.ajp.2024077
Chengappa, K. N. R., Gannon, J. M. and Joshi, Y. B. (2025) A Call to Psychiatrists: Deprescription of Unnecessary Anticholinergic Medications in Schizophrenia Must Start Now, American Journal of Psychiatry. 2025/04/01, American Psychiatric Publishing (AJP), 182(4), pp. 319–321. 10.1176/appi.ajp.20250124
