Article Text

Download PDFPDF

Managing acute behavioural disturbance in a neurology ward
  1. Alan Carson1,
  2. Tracy Ryan2
    1. Correspondence to Dr A Carson, Department of Clinical Neurosciences, Western General Hospital, Edinburgh EH4 2XU, UK; a.carson{at}

    Episodes of behavioural disturbance are commonplace on neurology wards. The key to their safe management is prior planning, including securing the safety of the ward environment, and a formal risk assessment. Protocols which define staff roles and the unit strategy for the management of behavioural disturbance should be formally agreed and appropriate training provided. The commonest cause of disturbed behaviour in neurology—and neurosurgical wards—is delirium. This presents with fluctuating orientation, grossly impaired attention and disruption of the sleep–wake cycle. The cause is generally multifactorial with a combination of pre-existing vulnerabilities and acute precipitants. Management reflects this and depends on a multifaceted approach to medical care, including basic supportive measures, minimising polypharmacy and promoting orientation, early mobilisation and nutritional status, as well as treating the underlying medical conditions. Antipsychotic medication has a specific treatment effect but never as the only approach to the management of the delirious patient.

    Statistics from

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Episodes of behavioural disturbance are commonplace on neurology wards. The key to their safe management is prior planning, including securing the safety of the ward environment, and a formal risk assessment. Protocols which define staff roles and the unit strategy for the management of behavioural disturbance should be formally agreed and appropriate training provided. The commonest cause of disturbed behaviour in neurology—and neurosurgical wards—is delirium. This presents with fluctuating orientation, grossly impaired attention and disruption of the sleep–wake cycle. The cause is generally multifactorial with a combination of pre-existing vulnerabilities and acute precipitants. Management reflects this and depends on a multifaceted approach to medical care, including basic supportive measures, minimising polypharmacy and promoting orientation, early mobilisation and nutritional status, as well as treating the underlying medical conditions. Antipsychotic medication has a specific treatment effect but never as the only approach to the management of the delirious patient.

    More years ago than I care to remember, when working as a surgical houseman, I was ambling across the lawn between the doctor's residence and the wards when the morning took an unusual twist from the normal routine: an older patient with Parkinson's disease appeared on the hospital roof, and proceeded, at surprising speed, to the edge. Even more startlingly, my colleague and lifelong friend, Bill, appeared in hot pursuit. A decent tackle betrayed Bill's prowess as a rugby flanker but the patient, oblivious to his Parkinson's, put up reasonable resistance. Yells for chlorpromazine were made from the roof top but none was available in the drug trolley; and the suggested 10 mg was probably inadequate. Bill, who like the editor of Practical Neurology, knew more about seamanship than psychopharmacology, next called for a rope. His bowlines held firm and the patient was secured to a chimney stack in an unorthodox but effective intervention until much needed help arrived. We all returned to the ward round for Bill to be duly chastised for lateness and dishevelment. This was the first of a number of incidents in which I learnt where things can go wrong in the management of acute behavioural disturbance. … Alan Carson.


    The management of delirium often takes place in an atmosphere of alarm and uncertainty. Careful planning, leading to a number of practical steps for management, will help keep staff calm (Wellcome Library, London).

    In this article, we hope to pass on some tips that may help to ensure such incidents are better dealt with.

    The five PS: proper preparation prevents poor performance

    Neurology wards regularly face episodes of behavioural disturbance. Most units have at least one substantive incident every 6–12 months, and minor problems are commonplace. The key stage in their management is advanced preparation. All too often the first thoughts given to management are in the midst of an emergency incident and by then it may be too late to correct the problems. The incident above did not demand intensive medical training or resources but rather a cheap lock to prevent the delirious patient getting on to the roof. Preparation for the management of behavioural disturbance starts with an examination of the security procedures in the ward.

    • Environmental precautions are important yet often not initiated. A professional risk assessment of the ward environment should be conducted as it is unlikely that many of the easily remedied hazards will be identified by the average clinician. For instance, it may come as a surprise that windows opening more than 100 mm are a major hazard (figure 1). Consideration should be given to other environmental measures, including clear strategies to aid orientation such as adequate lighting, signs and clocks, the design of door latches, nursing observation points and quiet areas with space where more disturbed patients can be safely nursed.

    • The major procedural security consideration is whether or not to have an alarm system for patients leaving the ward (figure 2). Such electronic tagging can be a very effective nursing aid because it allows for a less restrictive approach to confused patients which can in turn lead to fewer ‘flashpoints’ where aggressive behaviour may develop in response to frequent restrictions imposed on the patient's movement.

    • However, it is relational security that is key to the management of disturbed patients. Staff must appreciate their role in the management of disturbed patients. It is never helpful having staff members acting as ‘barrack room lawyers’ and holding discussions about whether or not it is their job to restrain patients while a colleague is holding desperately on to a confused patient exiting the hospital, or worse trying to climb out of a window. There is a duty on the part of the hospital to provide a safe environment for the care of patients and, in the case of a neuroscience centre, it is reasonable to expect that as a ‘system’ it can cope with the management of common, and therefore predictable, episodes of disturbed behaviour. For instance, consideration might be given to potential roles for psychiatrically trained nurses or security guards. What matters is that all staff know what their role is and are aware of their responsibilities—policy should not be made ‘on the hoof’ mid-incident. Emotions often run high at such times and therefore a written policy, defining staff roles and the unit strategy, for management of behavioural disturbance is recommended. In turn, staff must be provided with the appropriate training if they lack skills in managing disturbed patients.

    An expert risk assessment of the clinical environment should be conducted. A window opening more than 100 mm represents a major hazard for falls. AC only discovered this in a tragic incident when one of his patients, with an acquired brain injury, squeezed through a third fl oor window, opening 130mm; he still struggles tobelieve it was possible. Thankfully the patient survived and made a complete recovery from her spinal injuries.

    Mechanical restraints, such as straight jackets, are rightly confi ned to the pages of history (a). There are, however, helpful procedural security measures such as electronic tagging that can greatly assist patient care (b). (Patient in a strait-waistcoat, by Ambroise Tardieu, 1838. Used with permission from the Wellcome Library, London.)

    Other considerations

    It is helpful to have a protocol or guidance outlining the legal principles underpinning care and what staff may or may not do (table 1). Regulations vary from country to country and familiarisation with local rules may be helpful as part of an induction course for newly appointed staff.

    Within the UK, management of acute behavioural disturbance takes place under Common Law Principles. The recent incapacity legislation in England and Wales, and separately in Scotland, does not alter this in an emergency situation. Key judgments are listed

    Emergency sedation is a common response to significant incidents and should be prepared for. It carries significant risk; in particular because physical aggression can alter the ordinary pharmacokinetics of intramuscular drug administration.1 Wards are advised to have a protocol for emergency sedation (table 2) and staff must be aware that drug doses vary between patients—for example, a typical adult may require haloperidol 5–10 mg intramuscularly to achieve an adequate effect but a frail older patient may be over sedated with 0.5 mg haloperidol. Importantly, wards should ensure that the drug trolley always has an available supply of the drugs on their protocol, and be aware of the need for observation of vital and other signs after any emergency sedation.

    An example of an acute sedation protocol for adults over 65 years of age

    Managing an incident

    It is far better to try to avoid episodes of behavioural disturbance (see below) but even with the best of care it is inevitable that some incidents will still occur. Unlike the management of most medical conditions—‘diagnosis before treatment’—these often require some treatment to actually allow assessment to begin. The pattern is often of acute onset of disturbed behaviour, emergency management (sedation), assessment and then further management.

    Adequate staff numbers to calm and if necessary to restrain

    The first step is to ensure adequate staff numbers are available at the scene. Help should be summoned sooner rather than later and one should aim to have extra staff physically present before they are required by necessity. They should not become immediately involved, or confront the patient, but be quietly and unobtrusively available for emergency assistance if needed. Three or four members of staff should go to the patient. One staff member, preferably familiar to the patient, should take the lead, stand out in front and try to engage the patient, speaking calmly but clearly, using open body language and reinforcing what is said with non-verbal gestures. He or she should use frequent repetition to aid orientation. Confrontation or chastisement is seldom helpful, and tactics such as distraction, “why not come and get a cup of tea”, are usually more effective. The patient should be contained within one area but staff should not intrude on the patient's personal space but rather stay at the perimeter of the area allowing some degree of movement within the limits of safety. There is no need to rush this stage and, if the situation is contained, plenty of time should be allowed for everything to calm down. Oral sedatives, for example liquid haloperidol, should be offered. If the situation escalates and restraint is needed, this should be done by a minimum of three staff, each knowing their role, and medication, according to protocol, administered. Observation and review are mandatory in the aftermath. A formal debriefing should be held as soon as is convenient and points of good and bad practice noted.

    In major psychotic illness, hallucinations tend to be auditory whereas in delirium visual hallucinations are more common


    Inevitably, clinical assessment usually takes place after emergency sedation rather than before. The patient is then often drowsy and it may only be possible to ascertain limited diagnostic information, at least to begin with. The common differential diagnoses are delirium, dementia, acute psychotic illness and, more rarely, personality disturbance. Diagnostic information is usually a composite of information acquired from the case notes, amalgamated with witness descriptions during the incident and any information which can be gleaned from the patient.

    There are two immediate clinical questions.

    • What was the patient's cognitive state before the incident? Sometimes this is obvious—for example, if the admission was after traumatic brain injury. On other occasions this will be less clear. Questioning of relatives at the first available opportunity can be very useful. A tool such as the IQCODE (Informant Questionnaire on Cognitive Decline in the Older) allows a reliable measure of pre-existing dementia (table 3).2

    • Was the patient orientated at the time of the incident? This can be difficult to ascertain, particularly if the patient is currently heavily sedated, but can sometimes be determined by either direct questioning or by inference from their speech and behaviour during the event. Disorientation is a cardinal feature of delirium, and may be present in dementia, whereas orientation is usually maintained in acute psychosis even though the patient may appear perplexed.

    The IQCODE (Informant Questionnaire on Cognitive Decline in the Elderly)

    Some estimate of the speed of onset should be made.

    • Delirium usually presents with an acute onset and a fluctuating course.

    • Dementia by contrast has an insidious onset and is slowly progressive. However, it must be remembered that patients with dementia are at high risk of developing superimposed delirium.

    • Manic illness can present relatively acutely over a few days and be mistaken for encephalopathy.

    • Schizophrenia tends to be more insidious, with deterioration over 3–6 months, although the development of actual psychotic symptoms can be over 2–3 weeks.

    • Psychotic depression usually has an insidious onset with a period of low mood giving way to an empty apathy.

    A mental state examination should be conducted. A key feature is the degree to which the patient's behaviour was goal directed (suggesting psychotic illness) as opposed to random interactions with the environment (more suggestive of delirium or dementia). In delirium there is also marked disturbance of the sleep–wake cycle and both hyperactivity and hypoactivity are recognised. Mania is accompanied by motor overactivity whereas in schizophrenia motor activity is often less marked and in depressive psychosis the patient is usually apathetic.

    Speech and language can be significant. As well as the usual neurological disruptions, one should look for the pressured speech with disjointed linkage of thought structure, often accompanied by rhyming or punning, suggesting an acute manic illness. In schizophrenia illnesses, the patient can show disjointed linkage of thoughts and ideas, with so-called ‘knight’s move' thinking, accompanied by neologisms and ‘stock’ words imbued with personal significance. Although formal thought disorder can sound like dysphasia, the construction of individual sentences is usually more complete, but with poor linkage between concepts, whereas in dysphasia individual sentences may be incomprehensible.

    Hallucinations, where present, can offer diagnostic clues. In major psychotic illness, hallucinations tend to be auditory (although visual and somatic hallucinations do occur) whereas in delirium visual hallucinations are more common. One should remember that visual hallucinations are also a core symptom of Lewy body dementia and occur in a number of more benign conditions such as Charles Bonnet syndrome. Critically, hallucinations in major psychoses tend to be linked to complex and sustained delusional schema which are firmly held, involve pathologically false judgement and they have overwhelming personal significance to the patient. In manic illness they are grandiose, frequently involving special powers and abilities. In schizophrenia they are classically persecutory. In depression, delusions are mood congruent and generally nihilistic. By contrast, in delirium, delusions, although common and often persecutory, tend to be simple, fleeting and more understandable in terms of the immediate circumstances of the patient. Delusions are less common in dementia and where present tend to be simple.

    Mood assessment is helpful too. In delirium and schizophrenia it is usually fearful. By contrast, in mania there is a shallow veneer of euphoria often giving way to irritability. In depressive psychoses apathy and indifference are more common than actual sadness.

    Orientation. Delirium has a fluctuating course, with orientation often preserved by day but deteriorating at night. Serial formal measurement of orientation every 4 h can be helpful in confirming the diagnosis. This is simple to do but often omitted or, perhaps through embarrassment, not done properly. The patient must be formally asked the time, date, where they are and who they are and the results recorded. It is not adequate just to have a chat and if the patient seems ‘normal enough’ record them as orientated.

    Cognitive function. The neuropsychological hallmark of delirium is impaired attention, accompanied by disturbance all domains of memory and in executive functions. Commonly used bedside tests of attention are tasks such as reciting the months of the year in reverse order or serial 7s (subtracting down from 100 by 7 each time). Bedside cognitive batteries such as the Mini-Mental State Examination3 and the more extensive, and more helpful, Addenbroke's Cognitive Examination (ACE-R)4 are of assistance. For a neuroscience unit we would recommend ACE-R, possibly supplemented by the Frontal Assessment Battery5 and some bedside tests of praxis, such as mimicking unusual hand positions and miming everyday activities. Attention is usually somewhat disrupted in patients who have active major psychoses but not nearly to the extent of that in delirium. In psychoses, memory function and executive function are, comparatively, preserved. It has been suggested that visuoperceptual deficits are particularly pronounced in delirium and that bedside tests such as dot counting, incomplete shape and incomplete letter recognition can help distinguish delirium from mild to moderate dementia,6 although their administration can be impractical in disturbed patients (figure 3).

    Visuo-perceptual skills, such as incomplete letter recognition and dot counting, may be particularly poor in delirium compared with early dementia (source: Addenbrokes Cognitive Examination).

    Personality disorder. Patients usually lack any features of substantive psychopathology. Rather they tend to be sensitive to perceived slights, and have a ‘paranoid’ but not delusional world view. Patient reports of ‘hearing voices’ often take the form of pseudohallucinations; voices speaking within their head as opposed to a true false perception of voices speaking in external space. Mood tends to fluctuate from upset to irritability. Boredom and low self-esteem tend to prevail over symptoms of genuine anhedonia (global inability to perceive pleasure). Incidents tend to develop over a period of time and are usually understandable, if unjustifiable, in terms of causation.

    Personality disorder can be a difficult area for clinicians. When the diagnosis has been clearly made, related disturbed behaviour is a criminal not medical matter and should be dealt with via ward staff for minor incidents (ie, an explanation that the person's behaviour is unacceptable), or by hospital security and, if indicated, the police for more significant acts. We do not recommend that staff simply ‘put up’ with or medicalise violent or aggressive behaviour.

    However, such a ‘zero tolerance’ approach needs to be considered alongside the degree of diagnostic certainty. Unfortunately, there can be little certainty in this field, and regrettably each year there are malpractice suits against clinicians who have discharged patients from hospital whose difficult behaviour was the direct result of their neurological disease and not ‘badness’.

    It making such a judgement, as well as the features of mental state described above, it can be helpful to look at the following.

    • The consistency of their behaviour—is there evidence to suggest that the displayed behaviour was in keeping with the patient's normal behaviour or was it out of character? Personality disorders, by definition, refer to enduring patterns of behaviour that are inflexible and pervasive.

    • The context of the admission: are acute conditions such as head trauma likely?

    • The context of the behaviour: aggression in personality disorder is usually displayed in the context of being thwarted—that is, a staff member saying no over some matter; something going wrong or causing displeasure; or receiving information that the patient does not want to hear. It is much less common for it to occur ‘out of the blue’—that is, sudden onset while lying in bed in the middle of the night.

    • The presence of alcohol and substance misuse: although be careful, the patient may be going into delirium tremens.


    The diagnosis of delirium and major psychotic disorders is based almost exclusively on clinical assessment. The role of investigations is confined to the exclusion of potential differential diagnoses. In a small number of cases, where there is considerable clinical uncertainty, an EEG can be helpful and will show generalised diffuse slowing in the vast majority of cases of delirium and dementia7 but it is generally normal in the major psychotic illnesses.8

    The main role for investigations is in identifying the underlying causes of delirium (table 4) and in cases of suspected dementia (table 5). In some situations investigations will be needed urgently if there is reason to suspect that the underlying cause may be a medical emergency (such as hypoglycaemia) but usually the investigation of the cause of delirium can wait until the next working day.

    Standard investigations in delirium. In many situations in a neuroscience unit the likely cause will be known and investigations will not be required. The list below is suggested as a ‘standard’ screening battery only

    ‘Standard’ investigations in dementia


    Major psychotic illness, accompanied by behavioural disturbance, should be managed in psychiatric inpatient settings. Urgent transfer should be arranged as soon as it is safe to do so. The common exceptions are post-ictal psychoses, steroid induced psychoses and dopaminomimetic psychoses, which all may be better managed in neurology wards. In such circumstances many units will favour having arrangements with psychiatric departments for accessing nursing skills—for instance, a registered mental nurse to assist with special observations. The policy for such arrangements should be agreed as part of the ‘advanced preparation’ and not involve staff arguing over it in the middle of the night while feelings are running high.

    The management of dementia tends to depend on local services, but in most areas of the UK more disturbed patients will be transferred to old age psychiatry. By contrast, delirium, in all but exceptional cases, will usually be managed in the neurology ward. We will therefore concentrate on the management of delirium for practical neurologists.


    Clinical features

    Although delirium is described as a disorder of acute onset, there is often a prodromal phase with mild transient symptoms of fatigue, restlessness, irritability and mood disturbance, accompanied by disturbance of attention or perception and sometimes hypersensitivity to light or noise. Sleep–wake disturbance is central to the condition; fragmented, disturbed sleep is often accompanied by daytime somnolence. Patients describe realistic dream activity at night and ‘dream-like’ states by day. Perceptual disturbance ranges from derealisation-like micropsia and macropsia through illusions (such as the cord of a dressing gown appearing like a snake), to frank visual hallucinations varying from simple shapes to complex forms.

    There are three variants of delirium: hyperactive, hypoactive and mixed. It is the hyperactive form with autonomic arousal that is more distinctive but the hypoactive form with inactivity, and on occasions falls and incontinence, is more common.

    Cognitively, the hallmark is decreased ability to focus, shift or sustain attention. These rapid shifts in attention have consequences for memory registration leading to disorientation and memory deficits. There are however global cognitive deficits and executive dysfunction, and disorganised thought can be prominent. Disturbance of retrograde memory is common and confabulation occurs. The Confusion Assessment Method (table 6) is a reliable, sensitive (94%) and specific (89%) tool for the diagnosis of delirium.9

    The Confusion Assessment Method (CAM) Diagnostic Algorithm


    Delirium has many causes, usually involving the interaction between pre-existing vulnerability factors and exposure to one or more acute precipitants. The greater the vulnerability, the more innocuous the precipitant may be. Conversely, in less vulnerable patients the insult will be more noxious and often multiple. Common vulnerabilities include: old age, male sex, cognitive impairment, depression, immobility, history of falls, hearing and eyesight problems, dehydration, malnutrition, polypharmacy and chronic medical illness. The common precipitants include: infections, acute illness, drugs and surgery but, in vulnerable patients, can also include factors such as admission to a strange setting (ie, a hospital ward), pain, sleep disturbance, minor physical procedures and emotional distress.

    The pathophysiology is poorly understood. Currently, most attention is directed towards the role of neurotransmitters, particularly acetylcholine and dopamine, inflammation and the stress response. There is extensive evidence to implicate cholinergic deficiency.10 Central to this theory is the role of cholinergic transmission in modifying attention which can be conceptualised as having four components: arousal, sustained attention, divided attention and selective attention. Orientation depends on global attentional processes and so marked disruption of attention inevitably leads to disorientation. The maintenance of attention depends on two major neural systems11: the ascending reticular activating system which exerts ‘bottom-up’ modulation of cortical regions and a cortical ‘top-down’ regulation, particularly involving prefrontal, limbic and parietal regions. Additionally, there is domain specific attentional modulation for such stimuli as sounds, faces objects and memories (figures 4 and 5). Disruption of the top-down system tends to produce inattention and distractibility as, for example, after traumatic brain injury or stoke, whereas disruption of the ‘bottom-up’ system produces delirium.

    The three compartments of an attentional matrix based on the work of Mesulam (from Hodges JR. Cognitive assessment for clinicians, 2nd edn. Oxford: Oxford University Press, 2007, with permission). ARAS, ascending reticular activating system.

    The ascending reticular activating system: major stuctures and pathways involved in normal attentional processes (from Hodges JR. Cognitive assessment for clinicians, 2nd Edn. Oxford : Oxford University Press, 2007, with permission).

    Acetylcholine is the major neurotransmitter in the reticulothalamic pathway, and the excitatory amino acids, such as glutamate, are relevant to thalamocortical transmission. There are also dopaminergic projections from the raphe nucleus and noradrenergic projections from the locus coeruleus. There is strong evidence from both human and animal models correlating the extent of cholinergic deficiency with the severity of delirium.10 A number of mechanisms have been postulated to cause cholinergic deficiency involving both direct and indirect central nervous system insults.10 Direct insults include stroke or brain injury, which cause ischaemia and related global impairment of cerebral metabolism.12 13 There is a glutamate and cholinergic surge in response12 13 which is followed by decreased cholinergic synthesis14 and ultimately a cholinergic deficit15 causing delirium. By contrast, indirect mechanisms involve a range of environmental and medical stressors which lead to astrocytic and microglial activation,10 16 and proinflammatory cytokine release.10 It has been suggested that these alter blood–brain barrier permeability,17 increasing monoaminergic activity7 and reducing the number of cholinergic neurons,18 in turn also causing a cholinergic deficit.15


    The mainstay of management is good general medical care and supportive symptomatic relief. Clinical assessment should be aimed at identifying as many of the pre-existing vulnerabilities and immediate precipitants as possible. When identified, the causative factor should be treated appropriately.

    • Supportive symptomatic relief is critical to safe patient management yet often ignored in favour of easily prescribed but less effective drug treatments.

    • Supportive relief includes steps to deal with any hypoxia, pain, dehydration, malnutrition and constipation (figure 6). Early mobilisation is very helpful and the time spent lying in bed should be minimised.19

    • Steps should be taken to assist with orientation of the patient, including ensuring that he or she is using the correct spectacles and hearing aids, have large faced clocks clearly visible and consider using an orientation board at the bedside.

    • When staff interact with the patient they should remember that inattention leads to poor registration of memory; repeating short simple phrases will aid registration and is more likely to be reassuring than long, albeit kindly, discussions.

    • Discontinuation or minimisation of drugs with sedative or anticholinergic adverse effects is essential. The use of analgesia is a balance between efficacy in controlling pain against the effects of opiates on cognition; for instance, it has been suggested that substitution of gabapentin for opioid based analgesia in postoperative patients helps to reduce the occurrence of delirium.20

    In the management of delirium, clinicians often rush to medication but the key to success lies in high quality general medical care and managing the more ‘mundane’ aspects of the presentation, such as early mobilisation and constipation (Used with permission from the Wellcome Library, London).

    This style of multi-faceted approach reduces morbidity and saves money. It has been translated into care pathways such as Hospital Elder Life Programme (HELP) which translate theory into practice.21

    Antipsychotic drugs

    Drug treatment should be reserved for more severe cases and antipsychotics are the agent of choice—they reduce the degree and duration of delirium.22 There is no evidence to support any particular antipsychotic over another. Low dose (<3 mg/day) haloperidol in older people has equivalent adverse effects to risperidone and olanzapine; in higher doses (>4.5 mg/day) it has more extrapyramidal adverse effects. In physically active younger patients, a daily dose of approximately 5–10 mg is usually safe. In terminally ill patients there is evidence to recommend haloperidol over chlorpromazine.23

    However, concern over the use of antipsychotic medication in older patients was raised by Janssen (the manufacturer of risperidone), highlighting a possible association between risperidone and stroke. In 2004, the Committee of Safety of Medicines in the UK recommended avoiding the atypical antipsychotic drugs risperidone and olanzapine in patients with dementia. Several studies have subsequently investigated this24,,26 and did indeed demonstrate an increased risk of stroke in patients receiving antipsychotic drugs. However, there was no particular association with risperidone, and if anything the risk was higher with first generation drugs, in particular the phenothiazines (such as chlorpromazine). Although such data emphasise the need for caution and restraint in prescribing, such small increases in risk should be compared with the significant risks delirium poses in terms of morbidity, mortality and patient safety; these risks should not be regarded as a reason for not treating delirious patients, at least in the short term. The reality is that there is often no choice but to treat delirious patients with some form of sedative, and the evidence strongly favours antipsychotic medication.

    By contrast to worrying about the long term risk of stroke in treating delirium, the emphasis should be on daily monitoring of the much more likely adverse effects of antipsychotic medication, particularly sedation, postural hypotension, extrapyramidal problems and worsening of cognitive function, all of which can lead to an increased risk of falls and in some cases even worsen the delirium.

    We use haloperidol as a firstline agent with an initial dose of 0.5 mg twice daily in the frail or older and 5 mg twice daily in younger fitter patients. Whatever dose is used, the drug should be titrated to the patient's response and any antipsychotic medication should be reduced or discontinued as soon as possible. The need for regular review of drug dose cannot be over emphasised. The use of ‘as required’ medication should be avoided, if possible, as it often results in unnecessary dose escalation. Where extrapyramidal problems develop, these are generally better managed by reducing drug dose than prescribing anticholinergic medications, such as procyclidine, which tend to worsen delirium.


    The 2009 Cochrane Review27 of benzodiazepines in the treatment of delirium found little evidence but what evidence there was did not support their use. Furthermore, the 2004 review of delirium in terminal care23 noted that lorazepam caused excessive sedation. The exception is the management of delirium tremens where there is a specific role for benzodiazepines28 (a common regimen is chlordiazepoxide 20 mg four times a day reducing over 5 days) and antiepileptic drugs.29


    The putative role of impaired cholinergic activity has led to the attractive hypothesis that acetylcholinesterase inhibitors may offer not only symptomatic relief but actual treatment. Despite a number of enthusiastic early case reports,30,,32 the 2008 Cochrane Review33 found only one trial of donepezil, in a mere 15 patients, to be methodologically adequate. Perhaps unsurprisingly there was no difference between active treatment and placebo and so there is no current evidence to support its use. But despite the lack of evidence, acetylcholinesterase inhibitors are routine in some countries,34 and more appropriately powered studies are currently being conducted (Dautzenberg PL, personal correspondence 2009). In the UK these drugs are currently unlicensed for the treatment of delirium.

    However, there may be a particular role for acetylcholinesterase inhibitors in patients who have delirium in the context of Parkinson's disease dementia where complex visual hallucinations are commonplace.35 Therefore, as well as reducing dopamine agonists, catechol-O-methyl transferase inhibitors and L-dopa, an acetylcholinesterase inhibitor, may be warranted, particularly given the relative contraindication of antipsychotics and benzodiazepines. Clozapine is the drug of choice36 but the limitations on prescribing, in particular the need to be registered with the monitoring company, mean that many favour quetiapine37 because it is easier to use.

    The main adverse effects of acetylcholinesterase inhibitors are cholinergic: dizziness, insomnia, nausea, vomiting and diarrhoea. These are dose and time limited and usually resolve with reduction in dose and when treatment is continued for more than 3–4 weeks. Other relatively common adverse effects are cardiac arrhythmias, particularly bradycardia, gastric and duodenal ulcers.


    There are very few randomised trials on the prevention of delirium. The Cochrane review in 200738 identified six studies mainly conducted in the setting of elective surgery in the older. One found that a proactive geriatrician consultation before elective hip surgery led to a reduced incidence in the treatment group (odds ratio 0.48, 95% CI 0.23 to 0.98); 10 areas were addressed—oxygen delivery, fluid and electrolyte balance, pain control, polypharmacy, bowel and bladder function, environmental stimuli, nutrition, mobilisation, early detection of postoperative complications and treatment of agitated delirium where present. One trial of low dose haloperidol in at risk patients pre-elective hip surgery showed no benefit but did result in shorter less severe episodes of delirium.

    In terms of identification of ‘at risk’ patients, evaluation of executive function and intercurrent depressive illness39 40 appear more predictive than memory tests, particularly when cognitive impairment is less overt.


    The management of behavioural disturbance and, in particular, delirium can be summarised in the Boy Scout motto ‘be prepared’. The clinical assessment and subsequent management encapsulates most of the skills of good general medical practice in combination with some interesting cognitive neurology. Unfortunately, assessment and management often occur in an emotionally charged atmosphere in which everyone feels the patient should be someone else's problem. Rarely, however, does ‘someone else’ appear to take it over and this can be a source of impotent frustration. We suggest that with proper preparation, both of the individual patient and the clinical care system, not only will poor performance be prevented but patient care will be rewarding and hopefully fun too!


    This article was reviewed by Tony Pelosi, Lanarkshire, UK.



    • None

    • Commissioned, externally peer reviewed.

    Other content recommended for you