Exploring Antipsychotic Use for Delirium Management in Adults in Hospital, Sub-Acute Rehabilitation and Aged Care Settings: A Systematic Literature Review

Delirium is a serious neurocognitive disorder that is common in patients with acute medical illnesses, chronic medical conditions and cognitive impairment. Incidence rates vary across clinical settings, and have been reported as between 33% [1] and 80% [2] in intensive care units (ICUs), between 26% and 62% during hospitalisation in inpatient palliative care units [3], 26% in acute stroke settings [4] and 3–29% in medical in-patients [5]. Additionally, an Australian study identified approximately 15% [6] of residents living in aged care homes presented to hospital with symptoms of delirium. Moreover, a study in Canada in 12 nursing homes found the incidence was as high as 40% during a nursing home stay (on average 32 months) [7].

Delirium is a sudden change in cognition characterised by inattention, disorganised thinking and an altered level of consciousness, the course of which fluctuates [8]. Three subtypes are commonly described: hyperactive, hypoactive and mixed delirium in which those affected move between the hyperactive and hypoactive states. Psychomotor agitation, aggression and perceptual disturbances, such as paranoia and hallucinations [9, 10], are often associated with hyperactive and mixed delirium.

These behavioural manifestations [9] can interfere with provision of care to people with delirium in hospitals and residential aged care [11], especially if staff lack the requisite knowledge, skills and resources to prevent and de-escalate behaviours [12]. A recent qualitative descriptive study that examined doctors’ and nurses’ experiences (n = 42) of antipsychotic use for people with delirium in Australia [13] identified that concerns for safety of nurses and the patient with delirium were the most common reasons antipsychotics were administered. Nurses also reported being unable to complete clinical work unless the person with disruptive delirium symptoms was sedated [13]. Both doctors and nurses reported they chose to use antipsychotics in anticipation that behavioural symptoms could escalate [13]. In a similar study, interviews with 21 physicians, nurses and pharmacists in an ICU in Canada also found that reasons given for antipsychotic use were to ensure patient and staff safety, promote sleep and settle patients when staffing was inadequate, and workloads were high [14]. International guidelines allow for antipsychotic use when safety is a concern and non-pharmacological strategies have not been effective. However, in these two studies [13, 14] it was not clear whether non-pharmacological strategies were implemented prior to antipsychotic use. It is a grave concern that systems issues around staffing and workload are a driver for antipsychotic use. Antipsychotics can be considered a ‘chemical restraint’ when used to make patients more manageable and to serve the perceived best interests of healthcare providers rather than patients [15]. Further research is needed to fully understand the prescribing and use of antipsychotics for people with delirium.

The rapid sedating effects of antipsychotics contributes to their frequent use in the management of disruptive behaviours [13, 15]. In a study that examined the occurrence of delirium in a cohort of over 2 million hospital patients in the USA, 29% (n = 24,787) of patients with delirium were treated with antipsychotics [16]. Similarly, a retrospective chart review study of 212 acutely ill patients aged ≥65 years in a geriatric ward identified the prevalence of antipsychotic use to manage delirium was 58% (n = 123) with an additional 23.1% (n = 49) receiving a combination of an antipsychotic and lorazepam [17]. In a retrospective cross-sectional study of patients aged ≥65 years (n = 133) admitted to an acute geriatric care unit, antipsychotic use was prescribed to 74% (n = 73) of patients in the management of delirium symptoms [18].

There is limited evidence that antipsychotic use in any clinical setting is either effective for managing delirium or enhancing clinical outcomes. In a multi-centre, blinded, placebo-controlled trial of adults admitted to an ICU (n = 987), treatment with haloperidol did not increase the number of ‘days alive’ [19]. Additionally, a randomised double-blinded trial of patients (n = 1183) with acute respiratory failure or shock and hypoactive or hyperactive delirium in an ICU found that antipsychotics did not significantly alter the duration of delirium [20]. The REDUCE Randomized Clinical Trial, which investigated the prophylactic haloperidol use in critically ill adults (n = 1789) at high risk of delirium in 21 ICUs in the Netherlands, found no improvement in survival at 28 days (p = 0.93) [21]. A randomised clinical trial (RCT) in 11 Australian inpatient hospice and palliative care services (n = 247, mean age 74.9, standard deviation [SD] 9.8 years) investigated the efficacy of risperidone or haloperidol on delirium severity, extrapyramidal effects, sedation and survival [22]. Those in both the risperidone (0.73; 95% confidence interval [CI] 0.09–1.37; p = 0.03) and haloperidol (0.79; 95% CI 0.17–1.41; p = 0.01) groups experienced more extrapyramidal effects than those in the placebo group [22]. Moreover, those receiving haloperidol also had worse overall survival (hazard ratio, 1.73; 95% CI 1.20–2.50; p = 0.003). The authors concluded that individualised management of delirium precipitants with non-pharmacological strategies in the placebo group resulted in less severe symptoms and a shorter duration of symptoms than when antipsychotics were given [22]. A systematic review of delirium in surgical (mean age range 61–87 years) and mixed medical/surgical/ICU (mean age range 39–84 years) settings, which included 16 RCTs (2288 participants), two prospective cohort trials (458 participants) and one historical controlled trial (476 participants), showed that antipsychotic use did not decrease the duration of delirium, length of stay or mortality [23]. Similarly, a systematic review of the efficacy of antipsychotics in hospitalised non-ICU patients, which included nine randomised and quasi‐randomised trials (727 participants), found no reduction in delirium severity, symptom duration or rate of mortality [24]. Moreover, a systematic review of the benefits and harms of antipsychotic use to manage delirium in hospitalised adults, which examined 16 RCTs (1768 participants) and ten observational studies (3839 participants), concluded there was little evidence to support the routine use of antipsychotics for delirium and there was an increased risk for cardiac side effects [25].

Older people (aged ≥65 years) are at a heightened risk of delirium, and the use of antipsychotics in this population is linked to significant harm. The anticholinergic, extrapyramidal and sedative effects associated with antipsychotics increase the risk of falls [26]. Polypharmacy in older people also increases the likelihood of drug interactions, and side effects such as QT prolongation (leading to rapid heartbeats) and extrapyramidal symptoms (parkinsonism, akathisia and tardive dyskinesia), as well as being associated with an increased risk of stroke, pneumonia [26‐28] and death [26, 29].

Internationally, current evidence-based guidelines do not recommend routine antipsychotic use for the prevention or first-line management of delirium. The American Geriatrics Society’s ‘Beers Criteria for Potentially Inappropriate Medication Use in Older Adults’ [28, 30] strongly recommends, based on moderate-quality evidence, to avoid antipsychotic use for behavioural management unless non-pharmacologic interventions have failed and/or there is a high risk of the person harming themself or others. If used, they recommend only the lowest effective dose [28, 30]. Similarly, the ‘Screening Tool of Older Persons’ Prescriptions (STOPP)—Screening Tool to Alert to Right Treatment (START)’ criteria [31] recommends to cease antipsychotic use in older people unless other care avenues have failed. In Australia, the Delirium Clinical Care Standard states that non-pharmacological strategies are first-line treatment, and antipsychotics are not recommended [32]. In situations where non-drug strategies are not working and there is imminent risk of harm because of behavioural symptoms, only using the lowest appropriate dose for the shortest possible duration is recommended (ACSQHC, 2021:28). Despite an international consensus that antipsychotics are to be avoided in the treatment and management of delirium (see summary of guidelines in Table 1), there are concerns that their use persists in clinical practice.

A preliminary search found no existing systematic review reported on the prescribing and use of antipsychotics in delirium. Hence, there is a lack of synthesised evidence on the rates of use and prescribing patterns of antipsychotics for delirium management across healthcare settings. Therefore, this review aimed to describe the prevalence and patterns of antipsychotic use in adult delirium management, with a specific focus on the proportions of people with delirium given antipsychotics, the types of antipsychotics used, the dosages, including the amount, frequency and duration, and the indications for use.

This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [33].

This review aimed to describe antipsychotic use for delirium management in clinical practice. We examined how often people diagnosed with delirium were given antipsychotics, the types of antipsychotics they received, the doses of those medications and if the initiated antipsychotic was continued at discharge. Currently, no drug is approved to treat delirium by either the Australian Therapeutic Guidelines (Psychotropic) [hereafter referred to as the Australian Guideline] [46] or the US Food and Drug Administration. The Australian Guideline recommendations on antipsychotic use in delirium were identified by the authors of this review as the most comprehensive available and served as the benchmark in this discussion. Delirium guidelines specific to clinical settings are also discussed regarding the findings where appropriate. The key findings of this review were that contrary to recommendations in international guidelines, in everyday clinical practice, antipsychotics were used to manage delirium: (1) in a high proportion of patients; (2) off-label and in doses that exceeded recommendations for safe use; and (3) for longer than recommended. Each of these are discussed in turn in the following sections.

High prescribing rates of antipsychotics to manage delirium found in this review indicated that the gap between best practice and clinical practice remains. Given the recommendation by numerous international guidelines to avoid antipsychotic use, especially in older people [32, 47‐53], this finding is particularly concerning. All guidelines suggest antipsychotics only be used after first addressing contributory predisposing and precipitating factors, or when symptoms are unsafe or causing significant distress to the individual [54], and, that they only be used in low doses for short periods [46, 51, 55]. We speculate the reasons for prescribing antipsychotics are related in part to the clinical context in which delirium occurs.

Delirium is a complex fluctuating condition that often occurs in medically unwell patients in settings where staff and resources can be suboptimal, and this can influence clinicians’ decisions to prescribe and use antipsychotics [13, 14]. Eight studies in this review reported indications for prescribing antipsychotics for delirium, such as behavioural concerns, acute agitation, to decrease patient distress, or restrain unsafe behaviours, e.g. physical or verbal aggression [16, 38, 56‐61]. This suggests safety concerns were commonly a factor in clinicians’ decisions to use antipsychotics, although there was little evidence provided to explain how the clinical context influenced the decision. In addition to safety concerns, Tomlinson et al. [13] identified that nurses were concerned by the considerable amounts of time and attention needed to provide care for people with hyperactive or mixed delirium especially when they had high workloads and were time poor. Nurses wanted people with delirium to be calm and settled otherwise they “can’t do my job” properly [13]. Nurses’ also experienced stress when they struggled to manage aggressive behaviours and would pressure doctors into prescribing antipsychotics [13]. A lack of people and staff trained to deploy non-pharmacological interventions for delirium, which are recommended first-line management [9], may contribute to the use of antipsychotics [13, 15].

Evidence has shown that the addition of trained personnel (e.g. volunteers) in the care of patients with delirium can reduce antipsychotic use. A pilot study in Poland investigated the effectiveness of a volunteer-led multi-component non-pharmacological intervention on the reduction in outcomes including antipsychotic prescriptions in internal medicine patients (n = 130) [62]. The intervention, which included regular patient visits to support orientation, physical activity, nutrition, hydration and sleep quality [62], significantly reduced the initiation of antipsychotics (16.9% intervention [n = 65] versus 32.3% control [n = 65], p = 0.04) [62]. The benefit of non-pharmacological interventions has also been corroborated in a study from the USA that examined the effect of the Hospital Elder Life Program (HELP), which involved a personalised care plan delivered by trained volunteers [63], on a range of outcomes including the prevalence of newly prescribed antipsychotics to hospitalised patients aged ≥65 years (N = 12,281) [64]. Antipsychotics were ordered significantly less often in the HELP group (n = 1411) than in the standard care group (n = 10,807) [8.9% n = 125 vs 31.5% n = 3400, p < 0.001] [64]. Of the 31.5% (n = 3400) of patients in the standard care group who received an antipsychotic, 85.3% (n = 2899) received a first-generation antipsychotic (including chlorpromazine, haloperidol, perphenazine, prochlorperazine, fluphenazine) [64]. For patients who were diagnosed with delirium in the HELP group (2.7%, n = 38), approximately 50% (n = 19) were given an antipsychotic compared with 62.4% of patients (n = 918) with delirium in the standard care group. However, more research is needed, not only to explore the effects of non-pharmacological strategies on antipsychotic use, but also how evidence-based approaches to implementation could support clinicians to adhere to best practices.

In this review, consistent with previous research, we identified that amongst the wide range of antipsychotics used, haloperidol was the most frequent across settings. Findings from two systematic reviews indicate that available evidence does not support the use of haloperidol or other second-generation antipsychotics to treat the symptoms of delirium in older adults [23, 25]. Furthermore, there is mixed evidence regarding the efficacy of haloperidol, it has potential for serious side effects [55] and has been linked to adverse outcomes [65]. Additionally, in a recent review of the use of haloperidol in older people with delirium, the Medicines and Healthcare products Regulatory Agency [55] advised special caution is needed. Evidence is lacking as to why, despite emerging evidence, haloperidol continues to be so widely used in delirium management.

The dosages of antipsychotics reported in studies in this review varied across settings but consistently exceeded that recommended in guidelines. Guidelines across all clinical settings recommend antipsychotics be only used at the lowest appropriate dose for the shortest possible duration [32, 51, 55, 66]. The Australian Guideline recommends haloperidol only be used if clinically indicated and then be given as a single oral dose of 0.5 mg, or up to 1 mg for younger patients [46]. However, we found reported haloperidol doses across all settings far exceed guideline recommendations.

In addition to haloperidol, use of olanzapine also exceeded recommended doses. Although numerous studies did not report every individual dose, the high daily average dose reported suggests multiple doses were given in a single day. The Australian Guideline suggests that, if used, a single oral dose of olanzapine between 1.25 and 2.5 mg be given, and up to 5 mg for younger patients [46]. We found daily doses of olanzapine generally exceeded guideline recommendations. Notably, use of risperidone adhered more closely to the recommended guideline of a single oral 0.5-mg dose with up to 1 mg for younger patients [46]. Studies in this review reported doses that ranged from 0.5 mg to 1.7 mg. However, in the subgroup analysis of people aged ≥65 years, the average dose of 0.97 mg was higher than recommended. In addition to excessive dosages, these findings highlight a significant variation in the amounts of antipsychotics used for delirium management.

International guidelines for antipsychotic use in delirium management are consistent and vary little between settings or clinical contexts and as such should be followed as closely as possible in all patients with delirium. Clinical Practice Guidelines for Adults in the Intensive Care Unit [53] recommend not routinely using haloperidol or an atypical antipsychotic to treat delirium. The recommendation is based on evidence that typical antipsychotics (e.g. haloperidol) and atypical antipsychotics (e.g. quetiapine, ziprasidone) are not associated with a shorter duration of delirium or decreased mortality. They further state that patients who experience significant distress or are at risk of harm may benefit from short-term haloperidol or an atypical antipsychotic until symptoms resolve [53]. A multicentre blinded placebo-controlled trial of 987 patients admitted to the ICU [19] identified that treatment of delirium using haloperidol did not lead to a greater number of days alive. Furthermore, a randomised, double-blind, placebo-controlled trial of 566 ICU patients [20] found that neither haloperidol nor ziprasidone significantly altered the duration of delirium (odds ratio, 0.88 [95% CI 0.64–1.21]. Doses of haloperidol used in ICUs in studies included in this review were extremely high and well above the recommendations for use of haloperidol in the ICU Clinical Practice Guidelines [53].

Guidelines for delirium management at the end of life [52] recommend avoiding antipsychotics in terminal patients. If pharmacological treatment is needed, a short-term low-dose antipsychotic is recommended for symptoms of perceptual disturbance, to control severe agitation, or if there are safety concerns [52]. Our review identified that ~70% of patients in palliative care settings had delirium managed with antipsychotics, and the average daily haloperidol dose of 3.05 mg exceeds recommendations. Even though patients may be terminal, potential risks associated with antipsychotic use should still be considered. A recent Cochrane systematic review [67] that investigated four RCTs of drug therapies in terminally ill adults (n = 399) concluded that haloperidol may slightly worsen delirium symptoms (median 0.49, 95% CI 0.10–0.88; n = 123) and increase extrapyramidal adverse events (median 0.79, 95% CI 0.17–1.41; n = 123). Furthermore, antipsychotics can appear to reduce some delirium symptoms through their sedating action but do not treat the underlying pathology of delirium [52]. Adding an antipsychotic can change delirium from hyperactive to hypoactive, which is equally distressing for the patient [52]. Shared decision making between clinicians, patients and family caregivers prior to antipsychotic use could reduce the risk of inappropriate prescribing and dosing, and facilitate active decision making to consider discontinuation when clinically appropriate [52].

There were very few studies that reported the use of antipsychotics according to the subtype of delirium. It is not surprising that for patients with hyperactive delirium, who may experience increased agitation and combative or uncooperative behaviour [9], there was a trend towards higher rates of antipsychotic use. As antipsychotics can reduce agitation and psychotic symptoms, and have a sedating effect, they are recommended if the patient is severely distressed and at risk of harming themselves or others [46]. However, studies in this review highlighted that some patients with hypoactive delirium also received antipsychotics [18, 45]. Patients with hypoactive delirium have reduced motor activity, increased lethargy and are quiet and withdrawn, symptoms that do not align with recommendations to consider antipsychotic use for reasons of agitation, distress or safety concerns. Indeed, antipsychotics are not recommended for the management of patients with hypoactive delirium [68]. The use of antipsychotics in patients with hypoactive delirium should be further investigated and future research should look specifically at delirium subtypes.

Antipsychotics were given repeatedly and over several days to many patients in studies in this review. This is well beyond the Australian Guideline’s recommendation of “a single dose of an antipsychotic is usually adequate” [46]. Irrespective of whether the number of doses or days of treatment were described, all studies in this review reported treatment that exceeded recommendations for the use of antipsychotics in people with delirium in all reported settings. The study by Tomlinson et al. [13] offers a possible explanation. In this qualitative study, clinicians (doctors and nurses) described that they would continue antipsychotics in anticipation of worsening behaviours. This means that antipsychotics were prescribed pre-emptively, in the absence of behavioural symptoms, to reduce the likelihood of disruptive agitated behaviours [13]. Regular evaluation of continued treatment with antipsychotics is essential [15]. More research needs to be done to understand clinicians’ rationale for giving repeat doses and prolonged treatment: if the intention is chemical restraint or “delirium treatment”, then an important focus of future research will be to investigate what will support clinicians to adopt best practice and support the use of non-pharmacological strategies early and throughout care.

Continuation of newly initiated antipsychotics at discharge was reported in only four studies [18, 39, 40, 44]. The rationale for post-discharge continuation of antipsychotics was not disclosed in the studies. The findings of this review indicated that a large proportion of newly prescribed antipsychotics were continued at discharge. According to the Australian Guideline [46], antipsychotics should be ceased as soon as possible; however, a short course of regular low-dose antipsychotics may be appropriate if the patient experiences severe distress that persists beyond a day [46]. Persistent use of antipsychotics beyond transitions of care may lead to avoidable adverse effects. Our study highlights that opportunities exist to examine the transition of care practices, regarding antipsychotic continuation or discontinuation at discharge. Further research is needed to understand the reluctance clinicians may have in discontinuing antipsychotics for people with delirium in readiness for discharge.

This systematic review presented evidence that antipsychotics are overused in the management of delirium. Studies included in this review demonstrated that antipsychotics remain a common choice to manage symptoms of delirium. Notably, haloperidol was the most frequently administered antipsychotic, often in doses that exceeded the recommended therapeutic amount in all clinical settings. The discrepancy between clinical practice and established guidelines is evident and potentially impacts on the quality of care. Further research is needed to determine the impact of antipsychotics on people with delirium and how to support the clinical workforce to implement evidence-based recommendations for antipsychotic use.