Risperidone for the Treatment of Dementia-Related Psychosis: A Systematic Review and Meta-Analysis

Dementia is a substantial public health burden affecting millions of individuals worldwide. It is a term used to describe a decline in cognitive function, affecting memory, thinking, and daily activities, and is often accompanied by behavioral changes [1]. Alzheimer’s disease (AD) is the most common cause of dementia, accounting for up to 80% of all cases, and is characterized by a gradual decline in cognitive function and daily living skills. The exact cause of dementia is not yet fully understood but is thought to involve a complex interplay of genetic, environmental, and lifestyle factors [2, 3].

Psychosis is a common neuropsychiatric symptom in dementia, affecting more than half of patients at some point during the disorder [4]. The manifestations of psychosis are variable and include delusions, hallucinations, and behavioral disturbances associated with significant distress and caregiver burden [5]. Antipsychotic medications are the mainstay of treatment for psychosis, but their use in psychosis is controversial due to questions about efficacy and adherence [6]. Their use in elderly patients with dementia is also controversial due to the risk of severe side effects such as stroke, extrapyramidal symptoms (EPSs), and impaired cognition [7].

Risperidone is one of the atypical antipsychotics that has been studied for the treatment of psychosis in patients with dementia [8]. Risperidone is a second-generation antipsychotic that works by blocking serotonin 5‐HT₂ and dopamine D₂ receptors in the brain [9]. Several randomized controlled trials have evaluated the efficacy and safety of risperidone in the treatment of psychosis in patients with dementia [10, 11]. It has been shown to improve psychotic symptoms in dementia patients. Still, it has also been associated with several side effects that need to be carefully monitored, such as sedation, weight gain, and orthostatic hypotension [11]. More importantly, dementia patients receiving risperidone had a 3.7% higher risk of mortality (95% CI: 2.2%–5.3%; p < 0.01) with a number needed to harm of 27 (95% CI: 19–46) compared with, respectively, matched nonusers [12]. However, this was not proved in an earlier meta-analysis of 17 placebo-controlled clinical trials [13].

A review study reported that risperidone was more effective than a placebo in reducing the severity of psychotic symptoms in patients with dementia. However, the study also found that patients taking risperidone were more likely to experience side effects such as sedation, weight gain, and movement disorders [5]. In that regard, a study revealed that the treatment assigned to 18% of dementia patients who received risperidone was discontinued due to intolerability, with a statistical significance compared to the placebo group [14]. On the other hand, a randomized double-blind placebo-controlled trial conducted by De Deyn et al. [10] found that risperidone significantly improved behavioral symptoms, including psychosis, in patients with dementia, with a low incidence of side effects.

Despite these mixed results, risperidone continues to be used for the treatment of psychosis in patients with dementia, often in combination with other non-pharmacological interventions such as behavioral therapy and environmental modifications [15]. Thus, the decision to use risperidone or any antipsychotic medication in these patients should be based on careful consideration of the risks and benefits, with regular monitoring for adverse effects.

Systematic reviews and meta-analyses are essential research methods used in evidence-based practice to identify gaps in the existing knowledge and highlight areas where further research is needed [16]. Here, we conducted a systematic review and meta-analysis to address the effects of risperidone on the treatment of psychosis in patients with dementia.

This study was conducted ethically in accordance with the World Medical Association Declaration of Helsinki. To investigate the effects of risperidone in reducing psychotic symptoms in patients with dementia, we conducted electronic searches on Medline via PubMed, Scopus, Web of Science, Google Scholar, and PsychINFO from the inception until May 2024. Additionally, we explored relevant studies by examining the reference list of each paper we identified. We conducted additional searches by examining the references of one meta-analysis related to the field, aiming to identify relevant papers for possible inclusion [17]. The search strategy was as follows: (TITLE-ABS-KEY [dementia] AND TITLE-ABS-KEY [psychosis] AND TITLE-ABS KEY [risperidone]) AND (LIMIT-TO [DOCTYPE, “ar”] OR LIMIT-TO [DOCTYPE, “cp”] OR LIMIT-TO [DOCTYPE, “le”]) AND (LIMIT-TO [LANGUAGE, “English”]).

The study included individuals diagnosed with dementia who met either the DSM-IV or ICD-10 diagnostic criteria for dementia (AD, vascular dementia, mixed dementia, or Parkinson’s disease [PD]) or the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer’s Disease and Related Disorders Association (NINCDS-ADRDA) criteria for AD or dementia with Lewy bodies, and also had concurrent psychosis [18, 19]. Psychosis diagnosis was established using the Neuropsychiatric Inventory (NPI) [20], the Behavioral Pathology in Alzheimer’s Disease Rating Scale (BEHAVE-AD) [21], the Brief Psychiatric Rating Scale (BPRS) [22], or the Psychotomimetic States Inventory (PSI) [23] tests. Our meta-analysis only included full-text studies conducted on humans published in English. Studies conducted on non-human subjects, such as in vitro, in vivo, or in silico studies, were excluded and not further evaluated. Additionally, we only considered studies that use risperidone as a treatment. The study methodology adhered to the guidelines recommended by Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [24].

To ensure consistency and high quality, the abstracts were first checked against specific inclusion criteria. Further screening was then carried out using the pre-determined exclusion criteria, with the kappa measure used (κ = 1.00) to ensure 100% consistency between two independent coders [25]. In cases where data were unavailable, the authors were contacted to provide raw or continuous data. Relevant information such as baseline participant data, treatment duration, and psychosis scores was retrieved from each report to calculate the effect size (ES). Studies were disqualified if the necessary outcome data could not be computed, imputed, or provided by the authors. The primary outcome of this study was the effect of risperidone treatment on the alleviation of psychotic symptoms in patients with dementia.

Two reviewers independently assessed the risk of bias (RoB) in the studies included by using the Cochrane RoB assessment method to evaluate sequence generation, allocation concealment, blinding, selective reporting bias, and attrition bias [26]. We visually inspected the funnel plots looking for asymmetries as a method to evaluate the probability of publication bias. Furthermore, we employed trim and fill analysis to assess publication bias in the studies that were included.

We utilized a random-effect model in our meta-analytic systematic review to account for heterogeneity in ES estimations, which allows for population-level inferences and is more stringent than a fixed-effect model [28]. As the I² statistic is not capable of informing us of the extent to which the ES varies, we did not use it to report the heterogeneity of the data. Its intended purpose does not include determining the variability of the ES; therefore, it cannot give this information unless the I² value is zero. Instead, we used prediction intervals for this purpose [29, 30].

During the search process, 521 records were identified. After an initial screening of titles and abstracts, 200 publications were considered potentially relevant. Further analysis led to the exclusion of 173 studies, leaving a total of 27 studies for the meta-analysis. However, eleven of these studies did not contain the necessary data and were subsequently removed [8, 10, 11, 31‒36]. The inclusion process of the research is illustrated by the PRISMA flowchart displayed in Figure 1. Table 1 displays the characteristics of the studies. The meta-analysis focused on 2,311 patients with psychosis due to dementia.

The initial global nested analysis of all psychosis measures using CMA included each study once. It was found that treatment with risperidone decreased psychotic symptoms in patients who had both dementia and concomitant psychosis, with a SMD of 0.355 (95% CI: 0.170–0.541, p = 0.000), as seen in Figure 2a.

Six studies (n = 354) [22, 39, 45, 46, 48] reported the data for the effects of risperidone on cognitive function. The analysis showed that risperidone treatment deteriorated cognitive function in dementia patients with psychosis (as observed by a lower Mini-Mental State Examination [MMSE] score), with an SMD of −0.185 (95% CI: −0.349 to −0.020, p = 0.028). The heterogeneity of the data calculated based on the prediction interval analysis was found to be low.

The findings indicated that the combined, opposing associations between the use of SSRIs and the intensity of psychosis symptoms in dementia patients were not attributed to any specific trial, as the analysis that excluded each trial demonstrated. Rather, these results reflected an overall pattern (Fig. 2b).

Based on the RoB assessment tool, the articles were given a positive rating for overall quality. The potential pleiotropic effects were checked and confirmed by the results of the funnel plot analysis, which showed no significant differences between the estimates in Figures 3 and 4. Additionally, after conducting both Begg’s and Egger’s tests, it was found that no studies were excluded in the trim and fill analysis, and there was no significant proof of publication bias with p values of 0.053 and 0.245, respectively.

In this analysis, the mean ES was 0.36 with a 95% CI of 0.17–0.54. Based on Figure 5, the true ES in 95% of all comparable populations falls in the interval of −0.37 to 1.08. This revealed a high heterogeneity among the included publications as the prediction interval showed a wider range of expected treatment effects than CI. A statistically significant effect is present here as all values in the 95% CI are on the same side of the null, while the corresponding 95% prediction interval indicates the possibility of values on either side of the null.

The results of the meta-regression indicated that the impact of treatment was positively associated with treatment duration (slope p = 0.038). Risperidone dose was also positively linked to the impact of treatment (two studies were not included in this analysis because they did not report the mean dose of risperidone at the endpoint of the study, slope p = 0.000) (see Fig. 6a, b, respectively).

We were not able to analyze safety data across the studies because a great number of studies lacked the corresponding data. However, we could report the adverse events observed during the studies.

The results were surprisingly heterogeneous. Barak et al. [20] and Ellis et al. [41] reported two and one EPSs in their patients, respectively. Hallucination (one case) and muscle rigidity (14.7%) were reported by Rainer et al. [45]. Seven studies found no significant changes in the scores of the adverse event paired with the beginning of the study in the risperidone group [21‒23, 39, 43, 47, 49]. In the Cruz-Jentoft et al. [37] study, the Undersøgelser Side Effects Rating Scale for Clinicians (UKU-SERS-Clin) extrapyramidal subscale ratings indicating a temporary reduction in EPSs during the study, with significantly lower scores at month 3. On the other hand, De Deyn et al. [38] found that the frequency of combined EPS-related adverse effects was higher in the complete risperidone group, with a rate of 16.3%, compared to the placebo group’s rate of 11.6%, and this seemed to vary with the dose. The same scale showed no significant change in the risperidone group compared to the citalopram group in the Culo et al. [46] study. The results of Pollock et al. [43] and Scheinder et al. [14] studies were in line with the findings of De Deyn et al. [38] report. This is contradictory to the findings of Duran et al. [40] who showed a noteworthy amelioration in EPSs, with a marked reduction in the UKU EPS subscale score from 4.0 at the start to 2.4 in the sixth week and further down to 1.7 at the 12th week (p < 0.01 compared to baseline).

Cruz-Jentoft et al. [37], reported two cases of hypotension, one somnolence, and one paresthesia [37]. In the Mintzer et al. [42] study, patients who were given risperidone experienced a higher incidence of treatment-related side effects compared to those given a placebo, with somnolence being reported significantly more often in the risperidone group (16.2% vs. 4.6%). Teranishi et al. [48] reported fall with contusion, swallowing difficulty, stridor, and sudden death in the risperidone group (n = 1 each). The body weight gain, body mass index increase, and elevation of prolactin levels were reported in the risperidone group compared with placebo in the Scheinder et al. [14] study.

Psychotic symptoms are common in patients with dementia and can cause significant distress to patients and their caregivers [8]. Evidence suggests that approximately 30% of individuals with dementia develop psychotic symptoms [50]. Psychotic symptoms can worsen cognitive and functional impairments in these patients, and increase morbidity and mortality rates [51]. Unfortunately, no consensus exists on treating symptomatic psychosis in individuals with dementia, despite its high prevalence. Our meta-analysis found risperidone effective in treating psychotic symptoms in dementia patients with concurrent psychosis, but it deteriorated cognitive function. However, a definitive pooled analysis of safety and adverse events data was not possible due to insufficient data from several studies. A landmark study on the effectiveness of atypical antipsychotic medications in AD patients is the Clinical Antipsychotic Trials of Intervention Effectiveness-Alzheimer’s Disease (CATIE-AD) [14, 52]. After 12 weeks, the Clinical Global Impression of Change (CGIC) scale indicated no variation in the effects of any of the atypical antipsychotics studied in this study. However, research demonstrated that the benefits of atypical antipsychotic medications for the management of psychosis in AD patients were outweighed by their side effects [14]. Continuation of the trial until 36 weeks showed that antipsychotics could work better for certain symptoms including aggressiveness, anger, and paranoid thoughts. However, it did not seem like they enhance life quality, care requirements, or functionality [52].

The findings of our meta-analysis provided evidence for the efficacy of risperidone in the treatment of psychotic symptoms in patients with dementia and concurrent psychosis. In line with that, risperidone has a long history of off-label prescription for the management of the behavioral and psychological symptoms of dementia (BPSD), including psychosis [53]. However, the efficacy data available for the use of risperidone in the treatment of dementia-related psychosis (DRP) are contradictory [14]. In this regard, Barak et al. [20] compared the efficacy of 6-week risperidone (1 mg/day) with escitalopram treatment in BPSD in AD patients and found that both treatment groups exhibited a reduction in NPI total scores, although the decrease was more prominent in the group receiving risperidone. Notably, the positive effects of risperidone treatment became apparent sooner, which is advantageous in addressing distressing and time-sensitive symptoms. On the other hand, Deberdt et al. [39] administered 1 mg/day of risperidone for 10 weeks in patients with dementia (AD, vascular and mixed dementia) and demonstrated that patients experienced improvement during treatment with risperidone, but the responses were not significantly different from the responses observed among patients receiving placebo. The differences between the results of these studies could be due to the different target populations for which risperidone was administered. In the first study, the target group was AD patients but in the second survey, the focus was on mixed, vascular, and AD dementias. It has been shown that patients with distinct types of dementias respond differently to the same antipsychotic. For instance, Culo et al. [46] found that patients with dementia with Lewy bodies and psychosis show a poorer response to risperidone than those with AD and psychosis. Also, evidence suggests that the effects of risperidone on BPSD could be age and baseline MMSE score dependent. The differences in these measures among studies could also account for the variable responses observed in these studies [54]. Accordingly, this suggests that while risperidone can be effective, its benefits may vary depending on the specific characteristics of the patient group, including the type of dementia, age, and baseline cognitive function. Consequently, personalized approaches to treatment are essential to optimize therapeutic outcomes for patients with dementia and concurrent psychosis.

The effects of risperidone on cognitive function in dementia patients are also important. Our study showed that risperidone treatment deteriorated cognitive function in dementia patients with psychosis. Examining the literature reveals that while some studies show no significant effects of risperidone on cognitive function compared to placebo [38, 48], others report a minor improvement on the MMSE scale for patients treated with risperidone (SMD: 0.15; 95% CI [−0.39, 0.09]) [55]. There can be several factors that contribute to the variable effects of risperidone on cognitive function in patients with dementia-induced psychosis. First, individual differences in the severity of dementia may account for this variability. As dementia progresses, the cognitive decline becomes severe and global making it more difficult for medications to improve cognitive function [56]. Second, there is evidence to suggest that the duration of treatment with risperidone can impact its efficacy. One study found that antipsychotic treatment lasting longer than 6 months is associated with declines in cognitive function in patients with dementia [57]. Third, the dosage of risperidone can also influence cognitive function, with higher doses potentially causing more cognitive impairment. A meta-analysis of studies examining the use of antipsychotics in elderly patients with delirium found that antipsychotics are more likely to cause cognitive impairment at higher doses [58]. Lastly, individual patient characteristics and comorbidities can also influence the effect of risperidone on cognition. Patients with dementia often have multiple comorbidities, and medications used to treat those comorbidities can interact with the effects of risperidone on cognition. Also, atypical antipsychotics can lead to severe metabolic side effects including obesity, dyslipidemia, and type 2 diabetes. These metabolic disorders are associated with cognitive impairments themselves [56]. In summary, risperidone’s impact on cognitive function in dementia patients is mixed. Our study found cognitive decline, while some literature reports no significant effect or minor improvements. Variability is influenced by dementia severity, treatment duration, dosage, and patient characteristics, with advanced dementia, prolonged use, and higher doses linked to greater decline.

Unfortunately, we were not able to perform a definitive pooled analysis of the safety and adverse events data because several studies did not provide the required data for this purpose. The use of risperidone (and any other atypical antipsychotics) has become complicated due to newly identified safety concerns such as cerebrovascular adverse events (including stroke and transient ischemic attacks) and death [59, 60]. Evidence suggests that risperidone has the lowest mortality risk among antipsychotics that were used for the treatment of BPSD [51]. Nevertheless, a recent Cochrane review found that using risperidone raised the risk of death (RR: 1.36, 95% CI: 0.90–2.05, 17 studies, n = 5,032; moderately certain evidence) and major adverse events (RR: 1.32, 95% CI: 1.09–1.61, 15 studies, n = 4,316; moderately certain evidence), albeit these values were inaccurate [61]. Due to this, all second-generation antipsychotic drugs now come with a black-box warning from the Food and Drug Administration (FDA) regarding an increased risk of death in senior dementia patients [62]. To address this issue, it has been argued that the use of lower doses (0.25–2 mg/day) of risperidone over a short-term period of 6–12 weeks may result in a lower prevalence of safety concerns in the elderly population. In that light, the Australian Pharmaceutical Benefits Scheme recommended on January 1, 2020, that the use of risperidone be restricted to 12 weeks in severe AD [63]. It has been suggested that risperidone dose reduction should be performed in dementia patients based on their age and MMSE score. For those with moderate-stage AD who are 75 years old, clinicians should use a conservative approach and start with a low dose of 0.5–1 mg/day. However, in patients with severe AD who are also 75 years old, starting at a low dose of a maximum of 0.5 mg/day is recommended, and for those 85 years old, the dose should be halved [54]. Surprisingly, the impact of risperidone on BPSD in our meta-analysis was found to be dose- and duration-dependent meaning that higher doses and longer duration of treatment were associated with a better therapeutic response. In that light, the clinician should always judge the therapeutic efficacy of risperidone against its adverse events in each patient to find the optimal individual dose/duration of treatment. In conclusion, while incomplete data hindered definitive safety analysis in our meta-analysis, cautious dosing and short-term use are advised due to safety concerns like cerebrovascular events, prompting guidelines for age-based dose adjustments.

The most up-to-date and thorough evaluation of the impacts of risperidone on dementia-induced psychosis is presented in this meta-analysis. Also, this analysis involved more studies than meta-analysis that was published previously [17]. Despite these strengths, this meta-analysis also has certain limitations, such as the inclusion of articles that were only published in the English language causing a possible language bias. However, evidence suggests that the exclusion of non-English studies has no significant effect on the ES reported in the meta-analysis [64, 65]. Also, the results may have been impacted by including non-randomized controlled trials in the analyses. It has been found that non-randomized studies are more likely to be influenced by potential biases when evaluating interventions, so it is important to be cautious when including them in reviews and meta-analyses [44]. Besides, there is no gold standard for the diagnosis of psychosis in dementia. The existing psychosis measures differ in terms of materials that address psychotic symptoms and have not been developed especially for dementia. The possibility that the effect of risperidone on dementia-induced psychosis is scale-dependent is out of the question, and our data provide some evidence for it. In addition, our study included patients with various types of dementia. Different disorders may show variable responses to the same doses/durations of the same treatment. This may limit the generalization of the findings of this study to specific dementia types. Last but not least, we were not able to provide an analysis of the safety data because many studies had not reported them. This is important as the safety concerns surrounding risperidone use in dementia-induced psychosis hinder its use for this purpose.

There are very few options are available for treating DRP. In general, the findings suggest that the use of risperidone is associated with enhanced outcomes regarding psychosis in dementia patients. This effect was found to be dose and duration dependent. However, we found that risperidone treatment was linked to poorer cognitive outcomes in these patients and the heterogeneity of data was high. Given the safety risks associated with risperidone use in patients with dementia, along with the considerable variability in data across studies, it is advisable to conduct randomized, placebo-controlled clinical trials examining the effectiveness and safety of various dosages and durations of risperidone in different types of dementia before its routine use in treating DRP. Also, for better outcomes, clinicians should prioritize the implementation of the DICE (describe, investigate, create, and evaluate) approach before prescribing risperidone to patients with dementia-induced psychosis. By utilizing this approach, clinicians and caregivers can more successfully detect which patients may benefit from risperidone treatment.

A statement of ethics is not applicable because this study is based exclusively on published literature. This study has been solely performed on subjects from previously published studies and consent to participate statement is not applicable to this study.

All the authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest (such as personal or professional relationships, affiliations, knowledge, or beliefs) in the subject matter or materials discussed in this manuscript.

The authors declare that this study did not receive any funds/financial support from anywhere.

Conception and design: Guanghua Zhou and Suna Yin; analysis of data: Guanghua Zhou; drafting of the manuscript: Shubao Zhang, Fang Hao, and Suna Yin; and revision: Lin Ma. All authors contributed important intellectual content and read and approved the final manuscript.

Guanghua Zhou and Suna Yin contributed equally to this work.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.