Antipsychotics
| Type | Recommendation |
|---|---|
| When to deprescribe | |
| CBR |
We suggest deprescribing be offered to older people who are taking antipsychotics when:
|
| Ongoing treatment | |
| CBR |
BPSD If deprescribing is unsuccessful despite multiple attempts and non-pharmacological options have been considered, we suggest maintaining the lowest effective dose, provided this aligns with the individual preferences, goals and overall treatment plans, and that the benefit clearly outweighs the harm and that potential underlying causes for BPSD (e.g. pain, constipation, depression) have been considered and/or addressed. However, we suggest reassessing the need for long-term therapy periodically at least every three to six months or more frequently if symptoms change and additional monitoring for cardiometabolic risks in people using antipsychotics (weight, blood pressure, lipids, diabetes screening, cardiovascular risk calculation). |
| How to deprescribe | |
| CBR |
We suggest individualising the tapering schedule and adjusting it according to the individual's response. In general, we suggest reducing the dose by 25% to 50% every one to four weeks, ensuring the absence of withdrawal symptoms and/or symptoms indicative of relapse before initiating further tapering. Once half the lowest standard dose formulation is reached, we suggest ceasing completely. However, smaller dose reductions may be appropriate or preferred by some individuals, particularly as lower doses are approached. |
| GPS | Healthcare providers should consider and offer adequate non-pharmacological management options such as psychosocial practices, structured care protocols, and sensory practices during and after deprescribing (ungraded good practice statement). |
| GPS | Healthcare providers should consider and offer appropriate management strategies to the individual’s families and care providers (e.g. verbal de-escalation, psychological intervention, increased staff-to-patient ratio, increased staff training in behaviour management (ungraded good practice statement). |
| Monitoring | |
| CBR |
We suggest closely monitoring for individual responses and tolerance to antipsychotic tapering, paying specific attention to changes in psychological withdrawal effects (e.g. agitation, hallucinations, anxiety) and physical withdrawal symptoms (e.g. dyskinesia), function, and quality of life every one to two weeks following each dose adjustment until at least four weeks after the medicine is fully ceased if practical. After this initial period, we suggest monthly monitoring for at least three months, followed by monitoring every six months thereafter. However, this should be tailored based on individual factors such as their preferences, responses and tolerance to deprescribing. If in-person visits are impractical, we suggest advising people to report symptoms as needed. |
| GPS |
Healthcare providers should use validated assessment tools to evaluate changes in neuropsychiatric symptoms, functional status, and quality of life (e.g. Neuropsychiatric Inventory for neuropsychiatric symptoms, Functional Status Questionnaire for functional status, and EQ-5D for health-related quality of life) (ungraded good practice statement). |
CBR, consensus-based recommendation; GPS, good practice statement
Antipsychotics are primarily used to treat psychotic disorders, such as schizophrenia and bipolar disorder, by alleviating symptoms like hallucinations, delusions, and abnormal behaviours/thoughts [177]. Their sedative and tranquillising effects can also help manage severe aggression and behavioural disturbances [177]. As a result, they are sometimes prescribed off-label in older adults for conditions such as behavioural and psychological symptoms of dementia (BPSD), insomnia, and anxiety disorders [611]. However, antipsychotic use for non-psychotic conditions should generally be avoided due to significant adverse effects, some of which are severe and potentially irreversible. In older people with dementia, antipsychotic use is particularly concerning, as it is associated with an increased risk of mortality and cerebrovascular events [612].
Antipsychotics are sometimes used to manage severe agitation, aggression, or psychotic symptoms in dementia, but they pose a high risk of serious side effects and offer only modest symptom relief [613]. BPSD encompasses agitation, aggression, hallucinations, delusions, depression, wandering, disinhibition, and vocal disruptions (e.g. calling out and screaming) [614]. These symptoms can be distressing for both the individual and their caregivers. While BPSD typically occurs or worsens with dementia progression, underlying factors such as physical pain, constipation, fatigue, or loneliness can also precipitate BPSD and should be assessed as part of the management plan [615].
A 2016 systematic review found that atypical antipsychotics, such as risperidone, olanzapine, and aripiprazole, provided limited improvements, while quetiapine was less effective. Typical antipsychotics showed similar modest benefits [616]. Importantly, any potential benefits should be observed shortly after initiation, as a lack of response within two weeks suggests further improvement is unlikely [617]. Some symptom resolution may occur naturally over time, which can overstate the perceived effectiveness of antipsychotics [613].
Non-pharmacological interventions are recommended as first-line treatments for BPSD [618-620]. These include psychosocial therapies (e.g. validation therapy, reminiscence therapy, music therapy), structured care approaches (e.g. bathing, oral care routines), and sensory interventions (e.g. aromatherapy, massage, bright light therapy, pet therapy) [621]. When BPSD symptoms are particularly distressing or pose a safety risk, the Monash Clinical Practice Guidelines suggest short-term risperidone at the lowest effective dose for acute behavioural disturbances [622]. However, atypical antipsychotics are not recommended for symptoms such as vocal disruptions, wandering, or disinhibition due to their risks and lack of proven benefit [622].
Refer to the narrative evidence summary, the GRADE Summary of Findings table in the guidelines, and the Technical Report for a complete presentation of the deprescribing evidence based on the GRADE framework (including other factors considered in developing the recommendations).
Antipsychotics can cause orthostatic hypotension, confusion, anticholinergic effects (e.g. constipation, dry mouth, dry eyes), and extrapyramidal side effects (e.g. dystonia, akathisia, parkinsonism, tardive dyskinesia) [623]. Long-term use is associated with metabolic complications, including weight gain, diabetes, metabolic syndrome, and QT prolongation. Older people, particularly women, have an increased risk of tardive dyskinesia, which may persist even after discontinuation [177]. The use of antipsychotics in dementia increases the risk of all-cause mortality and stroke risk [624, 625].
Given these risks, deprescribing should be considered in cases where antipsychotics are prescribed inappropriately or where the indication is no longer relevant [623]. Factors such as adverse effects, drug interactions, high drug burden index (DBI), poor adherence, or patient preference should prompt discussions between healthcare providers and patients about deprescribing opportunities [623]. DBI is a pharmacological measure used to quantify an individual's exposure to medicines with anticholinergic and sedative properties, including antipsychotics [434]. A high DBI is associated with an increased risk of falls, cognitive decline, and reduced physical function [626].
When antipsychotic therapy is identified as suitable for deprescribing, the tapering plan should be individualised, ensuring the absence of physical or neuropsychiatric withdrawal symptoms before initiating further tapering. The general approach of tapering a long-term antipsychotic used for BPSD involves approximately reducing the dose by 25 to 50% every one to two weeks until the lowest practical dose is reached, then after one to two weeks, stop the antipsychotic [627]. Slower tapering may be necessary for people with a high baseline dose or those with a history of severe symptoms [627]. The Maudsley deprescribing guidelines recommend a hyperbolic tapering strategy, using progressively smaller dose reductions at lower doses, to reduce withdrawal risks and relapse [43]. This method may be preferred by some individuals seeking a gradual transition off antipsychotics.
We identified 18 studies (eight RCTs, one prospective cohort study, and nine before-and-after studies) related to antipsychotic deprescribing from the systematic review and meta-analysis [628-643].
Overall, the current evidence for deprescribing antipsychotics is derived from studies of low and very low certainty with a focus on people with cognitive impairment using antipsychotics for BPSD, especially in nursing home residents. There was no evidence for deprescribing for psychotic disorders except for one study that investigated drug-induced psychosis in people with idiopathic Parkinson’s disease [637]. It was also difficult to interpret the findings given the potential confounding factors related to the use of concomitant psychoactive drugs, including antidepressants, benzodiazepines, and sedative hypnotics. The majority of the studies have very short follow-up periods (one to 12 months) so it may be possible that severe and persistent withdrawal syndromes, if occurred, were not captured. There was also a lack of clear differentiation between withdrawal effects and relapse of the initial condition in many studies. The evidence at this stage is insufficient to inform evidence-based recommendations.
If antipsychotics are considered appropriate to deprescribe, closely monitoring for psychological withdrawal effects (e.g. agitation, hallucinations, anxiety) and physical withdrawal symptoms (e.g. dyskinesia), function, and quality of life may be appropriate.
Key study characteristics and results
A narrative summary of each study is provided below, highlighting key characteristics and main findings.
Ruths 2008 randomised 55 nursing home residents with dementia taking antipsychotics for BPSD to the placebo group (n=27) or continuation (n=28) [639]. Participants were included if they were taking either haloperidol, risperidone, or olanzapine for at least three months, and were excluded if it was prescribed primarily for a major psychotic disorder or if they had an intellectual disability. Standing orders for concomitant psychoactive medicines (e.g. antidepressants, hypnotics, and anxiolytics) were permitted during the study. At week four, 23/27 participants (85%) in the intervention group remained off antipsychotics. There was no significant difference between the two groups in terms of mortality (OR 3.38, 95% CI 0.33, 34.65), neuropsychiatric symptoms (MD 3.00, 95% CI 0.16, 5.84), or behavioural deterioration (OR 2.16, 95% CI 0.18, 25.32). In an earlier sub-analysis (Ruths 2004) involving an actigraphy assessment in 30 participants, discontinuation of antipsychotics significantly reduced sleep efficiency (sleep time was 54 minutes shorter in the placebo group; p=0.029) [644].
Van Reekum 2002 randomised 33 nursing home residents with dementia who were taking a stable dosage of antipsychotics for at least six months to discontinuation (n=17) or continuation (n=16) [642]. Participants were excluded if they had delirium, a history of schizophrenia, behavioural symptoms that were disturbing to caregivers in the past two weeks or were using antipsychotics for treating nausea only. On-demand use of 0.5 to 1.0mg lorazepam was permitted for agitation during the study. At six months, there was no significant difference between the two groups in terms of mortality (OR 0.44, 95% CI 0.04, 5.36) or behavioural deterioration leading to study withdrawal (OR 1.33, 95% CI 0.25, 7.17).
Bridges-Parlet 1997 randomised 36 nursing home residents with possible or probable Alzheimer's disease who were taking a stable dosage of antipsychotics for at least three months to discontinuation (n=22) or continuation (n=14) [633]. Participants were included if they had a history of physically aggressive behaviour and were excluded if they were using antipsychotics primarily for a psychiatric disorder or had an intellectual disability. Concomitant use of antidepressants was permitted during the study if the medicine doses had been stable. At week four, there was no significant difference between the two groups in terms of agitation (OR 13.54, 95% CI 0.16, 79.29) or the number of physically aggressive behaviour episodes (MD -3.23, 95% CI -8.19, 1.73).
Devanand 2011 randomised 20 outpatients with Alzheimer's disease who previously responded to haloperidol treatment to discontinuation (n=10) or continuation (n=10) [645]. Participants were included if they were diagnosed with dementia and probable Alzheimer's disease and had current symptoms of psychosis (i.e. presence of delusions and/or hallucinations) or agitation/aggression. Participants were excluded if they were diagnosed with a psychotic disorder, had delirium, alcohol or substance abuse or dependence in the past 12 months, had stroke, other dementia type, or movement disorders. Concomitant psychotropic medicines were not permitted during the study. Discontinuation of haloperidol was associated with an increased risk of relapse (8/10 [80%] on placebo relapsed compared to 4/10 [40%] on haloperidol). Relapse was defined as at least 50% worsening from psychotic or agitation/aggression symptoms and score on the Clinical Global Impressions Scale.
Devanand 2012 randomised 110 outpatients with Alzheimer's disease who previously responded to risperidone treatment to the placebo group for 32 weeks (n=40), or continuing risperidone for 32 weeks (n=32), or continuing risperidone for 16 weeks before taking placebo for 16 weeks (n=38) [636]. Participants were excluded if they had a history of stroke, transient ischemic attack, or uncontrolled AF. Stable doses of concomitant psychotropic medicines (e.g. anxiolytics, hypnotics, and antidepressants) were permitted during the study. On-demand lorazepam at 1 mg or less per day was permitted. At 32 weeks, there was no significant difference in mortality (OR 0.38, 95% CI 0.03, 4.44) between the group that continued with risperidone and the group that took a placebo for 32 weeks. At 16 weeks, the placebo group (n=40) had an increased risk of relapse of psychosis or agitation compared with the two groups who had continued taking risperidone for 16 weeks (n=70) (OR 3.07, 95% CI 1.37, 6.86).
Ballard 2004 randomised 100 nursing home residents with possible or probable Alzheimer's disease who had been taking antipsychotics for longer than three months with no severe behavioural disturbances to discontinuation (n=46) or continuation (n=54) [631]. At week four, there was no significant difference in mortality (OR 1.19, 95% CI 0.23, 6.18), neuropsychiatric symptoms measured using the Neuropsychiatric Index (NPI) (MD 3.0, 95% CI -3.69, 9.69), or quality of life (MD -0.53, 95% CI -1.42, 0.36) between the two groups.
In a subsequent larger RCT (n=165), Ballard 2008 included nursing home residents with possible or probable Alzheimer's disease who were taking thioridazine, chlorpromazine, haloperidol, trifluoperazine or risperidone for three months or longer for BPSD [630]. Participants were randomised to discontinuation (n=82) or continuation (n=83). At 12 months, there was no significant difference between the two groups in the neuropsychiatric symptoms using the NPI, extrapyramidal symptoms, severity of Parkinson’s disease measured using the modified unified Parkinson’s disease rating scale, activities of daily living measured using the Bristol ADL, cognition measured using the standardised Mini-Mental State Examination or using the Severe Impairment Battery (SIB). However, there was a significant decline in verbal fluency, measured using the Verbal Fluency Task, among people who continued their antipsychotics (MD -3.80, 95% CI -6.91, -0.69). In the longer-term follow-up of up to 54 months, Ballard 2009 reported deprescribing was associated with reduced mortality (OR 0.51, 95% CI 0.28, 0.96) [646].
Cohen Mansfield 1999 reported a double-blind crossover RCT that included 58 nursing home residents who were taking either haloperidol, thioridazine or lorazepam (antipsychotics or benzodiazepines) for agitation [635]. Participants were excluded if they were taking other antipsychotics or anxiolytics concurrently (except for low-dose trazodone hydrochloride for sleep), had uncontrolled blood glucose, or had a diagnosis of major affective disorder of schizophrenia. There were no significant differences between the discontinuation and continuation period in terms of psychiatric symptoms measured using the Brief Psychiatric Rating Scale, physical aggression measured using the Cohen-Mansfield Agitation Inventory, global clinical status measured using the Clinical Global Impression Scale, or cognition measured using the Mini-Mental Status Exam.
Somani 1996 conducted a non-randomised study that included 57 nursing home residents with dementia who had been taking antipsychotics for at least three months for BPSD [640]. Participants who were diagnosed with movement disorders, or schizophrenia, or were using beta-blockers or long-acting benzodiazepines concurrently were excluded. In the study, 17 participants had their antipsychotics gradually discontinued, 18 participants had dose changes, and 22 participants had continued taking their antipsychotics. There were no significant differences between the control group (n=22) and the intervention group/dose change group (n=35) in terms of falls, dyskinesias measured using the Dyskinesia Identification System Condensed User Scale Instrument, or disease exacerbation. However, compared to the control group, the intervention group/dose change group had a significantly higher incidence of withdrawal dyskinesia (9 vs. 0, OR 32.14, 95% CI 1.67, 617.16) and behavioural relapse (11 vs. 0, OR 21.12, 95% CI 1.18, 379.52). Behavioural relapse was defined as an increase in target behaviours after dose reduction that necessitated a return to at least the baseline dose.
Thapa 1994 reported a prospective cohort study using an educational program to reduce antipsychotic prescribing in nursing homes [641]. A total of 271 nursing home residents who were receiving antipsychotics at baseline were included in this cohort study (discontinuation group, n=64; continuation group, n=207). There were no significant differences between the two groups in terms of involuntary movements measured using the Abnormal Involuntary Movement Scale, behavioural problems, measured using the Nursing Home Behaviour Problem Scale, depression measured using the Geriatric Depression Scale, activities of daily living, measured using the Lawton's Physical Self-Maintenance Scale, and cognition measured using the Mini-Mental State Examination. However, discontinuation of antipsychotics was associated with a significant reduction in observer-rated psychiatric symptoms measured using a modified Brief Psychiatric Rating Scale with a higher score indicating more severe symptoms (MD -0.36, -0.59, -0.13).
Azermai 2013 conducted a before-and-after study to abruptly discontinue antipsychotics in 40 inpatients with dementia or cognitive impairment who had been using antipsychotics for at least a month explicitly for BPSD [628]. The majority (n=31, 78%) remained off antipsychotics after one month, with an improvement in neuropsychiatric symptoms measured using NPI (-5.7, p=0.003, lower score indicates less severe symptoms) although mild withdrawal symptoms (e.g. physical symptoms such as nausea, vomiting, diarrhoea and psychological symptoms such as agitation, insomnia, anxiety, hallucinations) were observed in 72% of the participants [628].
Brodaty 2018 reported a before-and-after study in which an antipsychotic deprescribing protocol was implemented in 23 nursing homes [634]. At 12 months, there was a substantial reduction in the number of participants prescribed antipsychotics (reduced by 81.7%, 95% CI 72.4, 89.0). Compared to the pre-intervention period, discontinuation of antipsychotics did not lead to a statistically significant difference in BPSD, falls, hospitalisations, or cognition.
Fernandez 2005 reported a before-and-after study that discontinued clozapine or quetiapine in six community-dwelling people with idiopathic Parkinson's disease and no ongoing psychosis [637]. Participants were included if they were currently taking antipsychotics for at least 6 months, had a history of drug-induced psychosis, and were excluded in the presence of dementia with Lewy bodies. Concomitant medicines for Parkinson's disease remained stable throughout the study. Five of the six participants (83%) reported psychosis episodes two weeks to two months following the end of each tapering period. However, there was no change in the Parkinson's Disease motor severity measured using the Unified Parkinson's Disease Rating Scale (44.5 vs. 43.8, p=0.36).
Horwitz 1995 conducted a before-and-after study to compare the discontinuation of antipsychotics among nursing home residents with dementia based on clinical judgment and mandate [638]. Deprescribing by mandate was more likely to fail compared with discontinuing at the discretion of physicians (50% vs 5%). Among residents who restarted antipsychotics, the most common reasons were increased verbal and physical aggression. Antipsychotic withdrawal did not lead to an improvement in neurological performance, functional status, or cognition.
Bach 2017 conducted a before-and-after study that involved a pharmacist-led chart review of antipsychotic use in 20 nursing home residents with dementia [629]. Residents with a diagnosis of schizophrenia, schizoaffective disorder, schizophreniform disorder, mood disorders (e.g. bipolar disorder), Tourette's disorder, or Huntington's disease were excluded. In the study, the most common indication for antipsychotic use was verbal or physical aggression (35%), followed by resisting care (30%), agitation and restlessness (15%), verbal outbursts (10%), wandering (5%), and hallucinations (5%). Overall, antipsychotic use was reduced from 28 residents (21%) to 19 residents (14%) in the nursing home throughout the seven months.
Bravo-Jose 2019 conducted a before-and-after study that targeted the use of antipsychotic drugs in 35 nursing home residents with dementia and behavioural disturbances [647]. Residents were included if they met one or more of the following criteria: 1) stable for six months, 2) without treatment modification for more than a year, 3) severe side effects to antipsychotics, 4) receiving a typical antipsychotic, 5) receiving two or more antipsychotics, or 6) advanced functional impairment and advanced dementia. Participants were excluded if they were using antipsychotics for delusions, or hallucinations, or previously had a psychiatric condition. Antipsychotics were successfully deprescribed in 28/35 (80%) participants living in long-term care institutions and reduced to the lowest effective dose in seven participants (20%). Deprescribing of antipsychotics did not lead to significant changes in neuropsychiatric symptoms measured using NPI (12.9 ± 12.8 to 13.8 ± 16.7, p=0.124).
Westbury 2018 conducted a before-and-after study using a multi-component intervention to reduce the use of antipsychotics and benzodiazepines in aged care facilities [643]. Participants who were receiving respite or end-stage palliative care and those who were prescribed antipsychotics or benzodiazepines for severe psychiatric disorders were excluded. At six months, 39% who were prescribed antipsychotics or benzodiazepines had their dose either reduced or ceased. There was no significant change in BPSD, social withdrawal, quality of life, agitation, or aggression following the discontinuation of antipsychotics.
Bergh and Engedal 2008 conducted a before-and-after study that included 23 nursing home residents with Alzheimer's disease or vascular dementia who had been taking antidepressants or antipsychotics for three months or longer [632]. Deprescribing was implemented for antidepressants (n=11) or antipsychotics (n=12). Participants were excluded if they had a severe psychiatric disorder or diabetes mellitus. At 24 weeks, discontinuation of antipsychotics was not associated with a significant change in cognition measured using Severe Impairment Battery (49.9 ± 35.2 to 60.3 ± 19.5), neuropsychiatric symptoms measured using NPI (33.4 ± 23.9 to 32.0 ± 30.9), depression symptoms measured using Cornell's Depression Scale (7.6 ± 5.8 to 6.7 ± 6.4) or both motor and non-motor symptoms measured using Unified Parkinson Disease Rating Scale (3.9 ± 2.8 to 2.8 ± 1.6).
Various methods were used for deprescribing in the included studies and tapering was the most common method. While there was no direct evidence that any particular method was associated with the greatest benefits and harms, dose tapering is likely more acceptable than abrupt cessation and helpful in determining the lowest effective dose for some people requiring dose reduction rather than complete cessation.
In two RCTs, the dose was halved for the first week, then quartered of the original dose in the second week, then ceased in the third week (n=53, very low certainty) [642, 645]. For people on 2 or 3mg haloperidol daily, 1mg was given for two weeks, then ceased. People on 0.5 to 1 mg ceased haloperidol abruptly (study=1, n=20, very low certainty) [645]. Antipsychotics were ceased abruptly in two studies (n=66, very low certainty) [633, 644], among which one study stated the dose was halved for one week before complete discontinuation if people were taking daily dose > 50mg chlorpromazine equivalent (n=30) [633]. In one RCT, antipsychotics were tapered for three weeks before complete discontinuation (n=58, very low certainty) [635] whereas in another study dose was tapered over one week (n=11, very low certainty) [632]. The method of deprescribing was not described in the other three RCTs (n=375, n not stated in one study) [630, 631, 636].
In non-randomised controlled trials (all very low certainty), antipsychotics were tapered at a rate of 25% of the daily dose each month for four months (based upon the availability of suitable dosage forms) with a goal of discontinuation after a maximum of four months (study=1, n=57) [640] whereas method was not described in another study (study=1, n=271) [641].
In one non-controlled trial (very low certainty), antipsychotics were ceased abruptly (n=40) [628] and the method was not described in two studies (n=136) [638, 643]. Antipsychotics were gradually tapered in other non-controlled trials, as summarised below:
- Individualised titration schedule over two to eight weeks (study=1,n=6) [637]
- Gradual dose reduction (study=1,n=20) [629]
- Gradual tapering according to a deprescribing guideline (study=1,n=35) [647]
- The dose was halved every two weeks until after two weeks on the minimum dose, then ceased completely (study=, n=139) [634]