Delirium is a syndrome characterised by an acute disturbance of attention and awareness which develops over a short time period and fluctuates in severity over the course of the day. It is commonly experienced during inpatient admission in the terminal phase of illness. It can cause symptoms such as agitation and hallucinations and is distressing for terminally ill people, their families and staff. Delirium may arise from any number of causes and treatment should aim to address these causes. When this is not possible, or treatment is unsuccessful, drug therapy to manage the symptoms may become necessary.
This is the second update of the review first published in 2004.Objectives
To evaluate the effectiveness and safety of drug therapies to manage delirium symptoms in terminally ill adults.Search methods
We searched CENTRAL, MEDLINE, Embase, CINAHL and PsycINFO from inception to July 2019, reference lists of retrieved papers, and online trial registries.Selection criteria
We included randomised controlled trials of drug therapies in any dose by any route, compared to another drug therapy, a non‐pharmacological approach, placebo, standard care or wait‐list control, for the management of delirium symptoms in terminally ill adults (18 years or older).Data collection and analysis
We independently screened citations, extracted data and assessed risk of bias. Primary outcomes were delirium symptoms; agitation score; adverse events. Secondary outcomes were: use of rescue medication; cognitive status; survival. We applied the GRADE approach to assess the overall quality of the evidence for each outcome and we include eight 'Summary of findings' tables.Main results
We included four studies (three new to this update), with 399 participants. Most participants had advanced cancer or advanced AIDS, and mild‐ to moderate‐severity delirium. Meta‐analysis was not possible because no two studies examined the same comparison. Each study was at high risk of bias for at least one criterion. Most evidence was low to very low quality, downgraded due to very serious study limitations, imprecision or because there were so few data. Most studies reported delirium symptoms; two reported agitation scores; three reported adverse events with data on extrapyramidal effects; and none reported serious adverse events.
1. Haloperidol versus placebo
There may be little to no difference between placebo and haloperidol in delirium symptoms within 24 hours (mean difference (MD) 0.34, 95% confidence interval (CI) −0.07 to 0.75; 133 participants). Haloperidol may slightly worsen delirium symptoms compared with placebo at 48 hours (MD 0.49, 95% CI 0.10 to 0.88; 123 participants with mild‐ to moderate‐severity delirium).
Haloperidol may reduce agitation slightly compared with placebo between 24 and 48 hours (MD −0.14, 95% −0.28 to −0.00; 123 participants with mild‐ to moderate‐severity delirium).
Haloperidol probably increases extrapyramidal adverse effects compared with placebo (MD 0.79, 95% CI 0.17 to 1.41; 123 participants with mild‐ to moderate‐severity delirium).
2. Haloperidol versus risperidone
There may be little to no difference in delirium symptoms with haloperidol compared with risperidone within 24 hours (MD −0.42, 95% CI −0.90 to 0.06; 126 participants) or 48 hours (MD −0.36, 95% CI −0.92 to 0.20; 106 participants with mild‐ to moderate‐severity delirium). Agitation scores and adverse events were not reported for this comparison.
3. Haloperidol versus olanzapine
We are uncertain whether haloperidol reduces delirium symptoms compared with olanzapine within 24 hours (MD 2.36, 95% CI −0.75 to 5.47; 28 participants) or 48 hours (MD 1.90, 95% CI −1.50 to 5.30, 24 participants). Agitation scores and adverse events were not reported for this comparison.
4. Risperidone versus placebo
Risperidone may slightly worsen delirium symptoms compared with placebo within 24 hours (MD 0.76, 95% CI 0.30 to 1.22; 129 participants); and at 48 hours (MD 0.85, 95% CI 0.32 to 1.38; 111 participants with mild‐ to moderate‐severity delirium).
There may be little to no difference in agitation with risperidone compared with placebo between 24 and 48 hours (MD −0.05, 95% CI −0.19 to 0.09; 111 participants with mild‐ to moderate‐severity delirium).
Risperidone may increase extrapyramidal adverse effects compared with placebo (MD 0.73 95% CI 0.09 to 1.37; 111 participants with mild‐ to moderate‐severity delirium).
5. Lorazepam plus haloperidol versus placebo plus haloperidol
We are uncertain whether lorazepam plus haloperidol compared with placebo plus haloperidol improves delirium symptoms within 24 hours (MD 2.10, 95% CI −1.00 to 5.20; 50 participants with moderate to severe delirium), reduces agitation within 24 hours (MD 1.90, 95% CI 0.90 to 2.80; 52 participants), or increases adverse events (RR 0.70, 95% CI −0.19 to 2.63; 31 participants with moderate to severe delirium).
6. Haloperidol versus chlorpromazine
We are uncertain whether haloperidol reduces delirium symptoms compared with chlorpromazine at 48 hours (MD 0.37, 95% CI −4.58 to 5.32; 24 participants). Agitation scores were not reported. We are uncertain whether haloperidol increases adverse events compared with chlorpromazine (MD 0.46, 95% CI −4.22 to 5.14; 24 participants).
7. Haloperidol versus lorazepam
We are uncertain whether haloperidol reduces delirium symptoms compared with lorazepam at 48 hours (MD −4.88, 95% CI −9.70 to 0.06; 17 participants). Agitation scores were not reported. We are uncertain whether haloperidol increases adverse events compared with lorazepam (MD −6.66, 95% CI −14.85 to 1.53; 17 participants).
8. Lorazepam versus chlorpromazine
We are uncertain whether lorazepam reduces delirium symptoms compared with chlorpromazine at 48 hours (MD 5.25, 95% CI 0.38 to 10.12; 19 participants), or increases adverse events (MD 7.12, 95% CI 1.08 to 15.32; 18 participants). Agitation scores were not reported.
Secondary outcomes: use of rescue medication, cognitive impairment, survival
There were insufficient data to draw conclusions or assess GRADE.Authors' conclusions
We found no high‐quality evidence to support or refute the use of drug therapy for delirium symptoms in terminally ill adults. We found low‐quality evidence that risperidone or haloperidol may slightly worsen delirium symptoms of mild to moderate severity for terminally ill people compared with placebo. We found moderate‐ to low‐quality evidence that haloperidol and risperidone may slightly increase extrapyramidal adverse events for people with mild‐ to moderate‐severity delirium. Given the small number of studies and participants on which current evidence is based, further research is essential.
Anne M Finucane, Louise Jones, Baptiste Leurent, Elizabeth L Sampson, Patrick Stone, Adrian Tookman, Bridget Candy
Plain language summary
Drug therapy for delirium in terminally ill adults
Delirium is common in people with a terminal illness. A person experiencing delirium may be confused, lack concentration, have disturbed patterns of sleep and waking, and experience hallucinations. Delirium can start suddenly and can cause distress to both the person and their family. Delirium may be caused by the underlying disease with which the person is affected, or as a side effect of drugs or other symptoms. Often it is not clear why a person has delirium. The multifaceted nature of delirium makes its management challenging. When it is not possible to identify the underlying cause, drug treatments are sometimes used to manage the symptoms.
The aim of this review was to find out what we know about the effectiveness and side effects of drugs in the management of delirium in adults with a terminal illness. For the purpose of this review, terminally ill adults includes anyone with an advanced progressive illness such as advanced cancer, advanced dementia or organ failure, as well as those receiving hospice and end‐of‐life care. We compared drug therapy with placebo (a substance with no known active effect), usual care, or any other drug or non‐drug treatment.
Our search to July 2019 found four trials, involving 399 adults in total. Participants had advanced cancer (three studies) or advanced AIDS (one study), and all had symptoms of delirium. The drugs evaluated were antipsychotics (three studies) or benzodiazepines (one study), compared to placebo or each other, on their own or in combination with another drug or placebo.
Most studies reported the outcomes we deemed most important: delirium symptoms, agitation, and adverse events (side effects).
It was not possible to combine the data from different studies due to a lack of similarity between them. We found low‐quality evidence that certain drugs (haloperidol and risperidone) may slightly worsen delirium symptoms for terminally ill adults with mild‐ to moderate‐severity delirium. We found moderate‐quality evidence that haloperidol probably slightly increases adverse side effects for people with mild‐ to moderate‐severity delirium.
Quality of the evidence
We rated the quality of the evidence from studies using four levels: very low, low, moderate, or high. Very low quality evidence means that we are very uncertain about the results. High‐quality evidence means that we are very confident in the results. We found no high‐quality evidence. This was due to the small number of people taking part, the number of people dropping out of the studies, and the small number of studies.
We found low‐quality evidence that, compared to placebo, drug therapy (specifically haloperidol and risperidone) may slightly worsen delirium symptoms in terminally ill people with delirium of mild to moderate severity. We found low‐ to moderate‐quality evidence that these drugs may slightly increase adverse side effects. Given the small numbers of studies and participants on which current evidence is based, further research is essential.
Anne M Finucane, Louise Jones, Baptiste Leurent, Elizabeth L Sampson, Patrick Stone, Adrian Tookman, Bridget Candy
Implications for practice For adults with a terminal illness
This review found no high‐quality evidence to support or refute the use of drug therapy for delirium symptoms in terminally ill adults. There is low‐quality evidence that drug therapy, compared with placebo, worsens delirium symptoms in terminally ill people with mild‐ to moderate‐severity delirium, but our confidence in the effect estimate is limited. The true effect may be substantially different from the estimate of the effect. There is moderate‐ to low‐quality evidence that drug therapy for mild‐ to moderate‐severity delirium may slightly increase extrapyramidal side effects compared to placebo, but the likelihood that the effect is substantially different is moderate to high.For clinicians
We found no evidence of benefit associated with haloperidol or risperidone for the management of delirium symptoms in terminally ill adults with mild‐ to moderate‐severity delirium. There is low‐quality evidence that haloperidol or risperidone may worsen delirium symptoms compared with placebo for adults with mild‐ to moderate‐severity delirium. There is moderate‐quality evidence that haloperidol for mild‐ to moderate‐severity delirium probably slightly increases extrapyramidal side effects compared to placebo; and low‐quality evidence that risperidone may slightly increase extrapyramidal side effects. There is very low quality evidence that lorazepam with haloperidol may be more effective than haloperidol alone for the management of agitation in people with severe delirium, but we are uncertain of this.
There remains insufficient evidence on the effectiveness and harms of drug therapies for delirium in terminally ill adults. The NICE 2010 guideline focuses on hospitalised adults, and specifically excludes recommendations for those receiving end of life care. Clinical guidelines recommend identifying reversible causes of delirium and using non‐pharmacological approaches before pharmacological management (Bush 2018; CCSMH 2010; CCSMH 2014; NHS Scotland 2014; SIGN 2019). This includes reviewing all medication and stopping non‐essential drugs, maintaining hydration, controlling pain, promoting good sleep patterns, re‐orientating the person frequently, improving oral nutrition and mobility, checking for opioid toxicity, checking for infection, constipation and urinary problems, and reviewing the full blood count and biochemistry.
If pharmacological intervention is essential to control symptoms, then the aims of management should be determined by the multi‐professional team and the person’s family or supporters. In adults with advanced cancer who are near to death, experiencing severe distress and suffering, and whose symptoms of delirium are not relieved by standard approaches, a clinician may consider following the European Society for Medical Oncology recommended framework for the use of sedation in palliative care (Cherny 2014).For policy makers
There is no high‐quality evidence to support or refute the use of drug therapy for delirium in terminally ill adults. There is low‐quality evidence that terminally ill people with mild‐ to moderate‐severity delirium do not benefit from risperidone or haloperidol, but our confidence in the effect estimate is limited and the true effect may be substantially different from the estimate of the effect.For funders of the intervention
There is no high‐quality evidence to support or refute the use of drug therapy for delirium in terminally ill adults. There is low‐quality evidence that terminally ill people with mild‐ to moderate‐severity delirium do not benefit from risperidone or haloperidol, but our confidence in the effect estimate is limited and the true effect may be substantially different from the estimate of the effect.
Implications for research For researchers
There is an urgent need for research to determine the effectiveness and harms of drug therapy for delirium in terminally ill adults. There is a requirement for research examining the effectiveness of drug therapy on different delirium sub‐types: hypoactive, hyperactive and mixed. We did not identify any study examining drug therapy for hypoactive delirium, even though this is the most common form of delirium experienced by terminally ill people (Meagher 2012). Research is also needed on the role of targeted PRN (as needed) use of antipsychotics as opposed to scheduled antipsychotics.
There is a need for evidence on the effectiveness of haloperidol compared with placebo and other drug therapies. Haloperidol is recommended as the first line of treatment for terminally ill people with delirium (NHS Scotland 2014; Twycross 2017); the evidence base underlying this recommendation is limited, however, and is drawn from studies of people who are not terminally ill. Given low‐quality evidence that haloperidol may worsen delirium symptoms amongst terminally ill people with mild to moderate delirium, evidence regarding the contexts in which haloperidol is effective is needed.
There is some evidence from other populations that second generation antipsychotics are as effective as haloperidol in managing delirium (Wang 2013), with fewer adverse events, and are recommended in some guidelines (CCSMH 2010; CCSMH 2014). We identified only two studies evaluating a second‐generation antipsychotic (Agar 2017; Lin 2008); additional studies are now warranted.
Further research is needed on the multimodality management of delirium, specifically the combination of non‐pharmacological and pharmacological interventions. This is important given that current guidelines advocate using non‐pharmacological interventions before and alongside pharmacological approaches to prevent and manage delirium (Bush 2018; NHS Scotland 2014; SIGN 2019). Further research to inform future guideline development regarding what combinations of interventions are most effective, when and for whom, is required.
Midazolam is recommended as a first line treatment for refractory agitated delirium in the context of palliative sedation in the last days or weeks of life (Bush 2017; Irwin 2013; Twycross 2017), but has not been specifically evaluated as an intervention to manage delirium. Clinical guidance recommends that midazolam is used alongside haloperidol to treat terminal agitation when anxiety is present (Twycross 2017); however no RCTs have examined this. Evidence on the effectiveness of midazolam is needed to clarify its role in delirium management and understand better when it is most, and least, effective.
In line with the WHO 2018, we adopted a broad definition of palliative care in searching for eligible studies. We identified three studies involving adults with advanced cancer, and one with AIDS. There is a need for studies involving participants with other advanced progressive conditions, in particular advanced dementia, given its prevalence (Etkind 2017), and the complexities associated with identifying and managing delirium in this population.
We identified only four RCTs, reflecting the challenges of conducting clinical trials involving terminally ill people, in particular when people are unable to consent to research participation. There is a growing body of literature suggesting that some terminally ill people welcome the opportunity to take part in research and benefit from doing so (Middlemiss 2015). Agar 2017 has also shown for the first time that an evaluation of drug therapy compared with placebo is possible in a terminally ill population. Studies including a placebo or best supportive care arm need to be prioritised. It is vital that terminally ill people with delirium are not excluded from participating in delirium research studies, whilst ensuring that appropriate safeguards are in place (Sweet 2014). People with delirium will usually be unable to consent themselves; consequently procedures for advance consent and consent‐by‐proxy are essential.
We considered network meta‐analysis (NMA). This would have allowed assessment of the relative effectiveness of several interventions synthesizing evidence across a network of randomised trials. A key assumption underpinning the validity of the approach is that there are no important differences in trials included other than the treatments being compared (Cipriani 2013). We deemed our data unsuitable for network meta‐analysis for several reasons including important differences in the clinical characteristics of participants across studies; differences in the drug administration; and the small number of studies resulting in sparse connections between each intervention. Consequently there was a high likelihood that network meta‐analysis would yield imprecise results and we decided that reporting the individual study results was more meaningful. We suggest that as further RCTs are conducted, authors of future updates and related systematic reviews revisit the potential of conducting network meta‐analysis to synthesize evidence and assess relative effectiveness of different interventions across a network of studies.For design
Attrition rates in the included studies and the relatively small numbers of eligible participants in any one palliative care treatment unit suggest that future studies should involve participants recruited from multiple centres. Further efforts should be made to limit attrition, and outcomes should be collected even when participants stop receiving the allocated treatment.For measurement
Future research should carefully consider which outcome measures are most important to terminally ill patients and their families and carers. For instance, outcomes such as patient comfort or patient and carer distress may become more important than resolution of delirium symptoms as end of life approaches. Such outcomes need to be identified, validated and prioritised for assessment alongside delirium severity using valid scales.
The Nu‐DESC has been validated, but the composite score based on three Nu‐DESC items has not been validated, and focuses on assessment of distressing delirium symptoms. This sub‐scale should be validated before it is used as a primary outcome measure in future studies. Specific outcome measures to assess distress related to hypoactive delirium may also be required.
Survival was assessed in two studies, yet death is expected in terminally ill patients, and it is difficult to determine whether survival/death is associated with the intervention, or a natural occurrence given the study population. Further clarity on how survival outcomes are to be interpreted are required in future studies. A core outcome set for delirium studies is warranted.Get full text at The Cochrane Library
Evidence Central is an integrated web and mobile solution that helps clinicians quickly answer etiology, diagnosis, treatment, and prognosis questions using the latest evidence-based research. Complete Product Information.