The management of terminal delirium
Delirium is a distressing and disturbing clinical event. Palliation of the symptoms by multi-component interventions can be effective. The goal of interventions is to raise the deliriant threshold by combined symptom relief, environmental, psychological and pharmacological interventions. Haloperidol remains the drug of choice for delirium. For intractable delirious symptoms at the end of life terminal sedation may be indicated.
Keywords: Delirium, deliriant threshold, haloperidol, terminal sedation.
Delirium is a common, if not inevitable, syndrome during the final 24-48 hours of life of the terminally ill cancer patient. Prevalence figures of up to 88% are quoted. Delirium is a harbinger of death. To suffer delirium is frightening and distressing to patient, relatives and staff. It is frequently under-recognised and mis-treated. The causes of delirium in the terminally ill [Table - 1] are unlikely to be reversible. Generally several aetiological influences are operating simultaneously. The symptoms may be ameliorated and palliated despite it being symptomatic of severe multi-organ failure. The same meticulous assessment and care needs to be accorded to delirium as is to other symptoms in palliative care.
The DSM-IV criteria for delirium [Table - 2] are those of a rapidly emerging disturbance of consciousness and a change in cognitions, with fluctuating symptoms and evidence of organic aetiology. Delirium lies on the continuum between full consciousness and deep coma. Delirium is brain failure. The inter-relationship of precipitating and predisposing factors determines the clinical state [Table - 3]. The absolute strength and activity of the noxious agents precipitating the syndrome and the vulnerability of the host to such influences are the critical factors. If vulnerability at baseline is low, patients may be resistant to delirium despite exposure to significant precipitating stressors, but if vulnerability at baseline is high, delirium is likely to occur with exposure to relatively minor precipitating factors. Environmental influences may moderate these risks by muting the impact of the toxin and/or by strengthening the host. Individual differences in susceptibility to delirium induced by any given cause have been demonstrated. Conceptualising delirium in terms of a 'threshold' model, a well established model for epilepsy, is useful in understanding the various risk factors. The 'deliriant threshold' for each individual is unique. The risk factors are multiplicative rather than additive. Limited 'brain reserve', as in the dementing or immature infant brain, enhances risks as does the severity of the physiological insult. Previous delirium increases the risk more significantly than does age and pre-existing cognitive impairment, suggesting a kindling effect. We are all prone to delirium, it is a syndrome of universal susceptibility.
Principles of management
The cardinal rules of management of delirium were formulated in antiquity and with the exceptions of abandoning blood letting and the introduction of antipsychotic medications, these remain applicable [Table - 4]. Diagnosis and management occur concurrently. Reversal of aetiology may not be possible in the terminally ill. The goal of intervention is to provide symptomatic relief and to lower the deliriant threshold.
The primary prevention of delirium is probably the most effective treatment strategy, though not possible in the terminally ill. Education of staff to ensure prompt recognition of delirium and multi-component intervention strategies are effective measures reducing the duration and diminishing the severity of delirium. Identification of the risk factors for delirium and multi-component strategies such as orientation, sleep hygiene, noise reduction, mobilisation, hydration, oxygenation, rationalising medications and hearing and visual aids has been shown to reduce the duration of delirium. Whether these strategies are truly preventive or prevent symptom escalation of sub-syndromal delirium is uncertain. The height of the 'deliriant threshold' is maintained by such interventions.
The role of cytokines in delirium or, more particularly, the possibility that somatostatin and IGF-I (insulin-like growth factor) may prevent their rise, has resulted in them being postulated as potential neuroprotective agents against delirium. Physical exercise is known to increase IGF-I uptake by the brain, thus supporting the purported protective role of exercise in delirium and suggesting an adaptive purpose of the hyperactivity of delirium. Prophylactic antipsychotic medication in patients at high risk, at least anecdotally, appears beneficial. Prophylactic haloperidol, given pre-operatively and continued post-operatively, has been shown to shorten the duration of delirium, but not its incidence. With the advent of neurologically tolerable atypical neuroleptics such as quetiapine this practice deserves further study. The preventive use of cholinesterase inhibitors similarly has appeal.
Delirium is a medical emergency. The first priority is to minimise the risk of injury to the patient, other patients and staff. The symptoms of delirium are most containable in the familiar home environment. Informed consent for urgent interventions is often not obtainable, unless a lucid period presents. Urgent, indeed emergency and life saving, interventions are governed by common law doctrine - that is treatment may be given without informed consent if the treatment is such that medical colleagues would generally consider it appropriate and a reasonable person would want it. Physical restraint tends to increase agitation and enhance the risks of deep vein thrombosis and myocardial infarction. If unavoidable it should only be a very brief emergency procedure. Bed rails merely increase the height of the potential fall, without reducing the risks. Placing the patient's mattress on the floor is a safe, but admittedly undignified option. Wandering may be managed by "collusion" i.e. walking with the patient and re-directing them. Avoiding unnecessary confrontations, protecting the patient's personal space, avoiding "eyeballing" (by glancing and not staring) all enhance the patient's tolerance to interventions and calm their distress and irritation [Table - 5]. Observations need to be frequent and discrete.
Treatment of causes and symptom relief
Though up to 50% reversal rates of delirium have been reported in palliative care practice, such response rates are more probable early in the disease course. Particularly toward the last few days of life reversal is less likely, for disease burden is considerable. Even if medical and surgical treatments are curative, the resolution of delirious symptoms may still take several days or longer. Hyperactive delirium responds faster than hypoactive delirium. However some interventions even at a late stage may be very effective. Antibiotic treatment of a minor infection, opioid rotation and adequate analgesia are such examples [Table - 6]. Even when the aetiology cannot be reversed, interventions that partially ameliorate the causes should decrease the intensity of the delirium.
Therapeutic manipulations of the environment are under-utilised interventions and free from adverse effects. Noisy, unfamiliar and unwelcome perceptual stimuli further compound the existing impairment of sensory filtering in the reticular activating system (RAS) of delirious patients, thus amplifying delirious symptoms. A failing brain, like an ailing heart, benefits from the reduction of its workload, hence the desirability to restrict, but possibly not deprive, the sensory input. A moderate sensory balance and an unambiguous environment are preferable to sensory over-stimulation. Defective perceptual discrimination facilitates the occurrence of distortions and misinterpretations of sensory stimuli in busy environments. Ambiguous lighting provokes illusions and even hallucinations, thus the room should either be clearly lit or emphatically dark. Generally light is the preferable environment for delirium and certainly in those in whom visual misperceptions are common the room should be well lit. The noise level should be below 45 decibels during the day and less than 20 decibels at night. Limiting staff interactions with the patient, minimising staff changes and restricting visitors are important. Constancy of staff may be difficult to achieve and the theoretical advantages of nursing patients in quiet side rooms must be weighed against the risk of nursing neglect. The presence of personal artifacts and objects may reduce the unfamiliarity of a hospice environment. Temperature regulation may be important with respect to the sundowning phenomenon. A later acrophase of high core body temperature is implicated in the sundowner effect in Alzheimer's patients. The room temperature should be kept between 21.1 degrees Celsius and 23.8 degrees Celsius. A bland and simple environment is not a sensory deprived environment, but rather a sensory restricted one. It can reduce the perceptual load on the delirious brain, thereby easing its failure.
These are under-utilised, yet simple and effective. Nursing staff and supportive relatives are best placed to use these strategies. Simple and firm communication taking advantage of any lucid periods, focused on orientation, reassurance, explanation and clarification needs to be repeated often. Fear and bewilderment do often respond to these strategies, albeit transiently. Delirious persons have an inherent fear of losing sanity even though the symptoms may be temporary. Support to foster the individual's sense of control and competency are crucial. Some relatives, because of their own fears and distress, can aggravate, rather than ameliorate, disturbance.
If the above measures are insufficient to achieve symptomatic and behavioural relief the introduction of pharmacotherapy is required to augment these interventions. This decision should not be influenced by the relative's wishes, time restraints or communication difficulties between medical and nursing staff. The benefits and risks of pharmacotherapy need to be carefully considered. Medications are more likely to be prescribed in frenetic, hyperactive delirium because the disruption caused to others tends to demand intervention, though the individual's distress in hypoactive delirium is no less disturbing.
The primary intent of pharmacotherapy is to dampen the intensity of the core symptom - the impairment of consciousness - by reducing the "workload" of the RAS. Neuroleptics, antipsychotics or major tranquillisers (different terminology for the same class of medicines) calm, tranquillise or soothe by enhancing inhibition of sensory input, allowing improved filtering and processing of internal and external stimuli. Drugs are aimed at controlling agitated behaviours, psychosis and cognitive dysfunction, sometimes with the total amelioration of symptoms. The intent is not to sedate the patient - to immobilise, quieten or induce sleep- but rather to relieve symptoms.
Haloperidol is the drug of choice. The only double blind randomised trial conducted in delirium compared haloperidol, chlorpromazine and a benzodiazepine and proved the effectiveness of the antipsychotics in controlling the symptoms and the aggravating effect of lorazepam. Haloperidol is a versatile medication, readily administered by most routes and with dosage forms allowing accurate titration [Table - 8]. Low dose oral haloperidol (1-10 mg / day, less in the elderly) improves symptoms in most patients. Haloperidol has no absolute contraindications. It has a very wide margin of safety. Haloperidol is much better tolerated if administered parenterally. The major acute neurological side effects (oculogyric crisis, dystonia) predominately occur in young adults and the chronic neurological side effects (akathisia, Parkinsonism More Details) in the elderly and in the mid-dose range (5 - 20 mgs per day). Benztropine and/or a benzodiazepine and a reduction of the dose, are occasionally necessary if these adverse effects occur, however anticholinergic agents risk aggravating the confusion. Long term neurological side effects, such as tardive dyskinesia, are unlikely because of the relatively short duration of use for this indication. The major cardiological side effect of concern is the prolongation of the QTc interval and torsades de pointes. This is more likely in high dosages and its dangerousness is dependent upon pre-existing prolongation. Droperidol, another butyrophenone, is no longer recommended in view of its greater pro-arrythmic risks. Phenothiazines may be used but may cause severe postural hypotension, anticholinergic side effects (dry mouth, urinary retention, blurred vision, constipation, enhanced confusion), are painful or locally irritant if administered parenterally and induce significant but unpredictable sedation. Thioridazine causes unacceptable cardiac side effects and is no longer recommended. Levomepromazine and pericyazine are the phenothiazines most suitable, if additional sedation is required.
The new atypical neuroleptics are increasingly being trialled because of their better tolerability. There are reports of olanzapine aggravating delirium. Risperidone and quetiapine show most promise though short acting parenteral forms of these medications are needed.,
There are relative contra-indications to haloperidol in delirium. These are hepatic encephalopathy, delirium tremens (DT's), AIDS and Lewy body dementia More Details. Benzodiazepines are advised in these conditions, though they may aggravate delirium if administered in inadequate dosage to settle and sedate. In drug withdrawal states the anticonvulsant properties of benzodiazepines are required. AIDS and Lewy body dementia patients are exquisitely sensitive to the neurological side effects of antipsychotic medications, thus benzodiazepines are preferable.
If behavioural control is not achieved with antipsychotics, then sedation may be required. Sedation merely enhances the depth of consciousness impairment and if sleep is not achieved delirium is aggravated. At low dosage benzodiazepines relieve anxiety and at higher doses they induce sleep and deep sedation. Careful titration to effect is required. In combination with antipsychotic medication augmentation of effectiveness in delirium has been proposed, but this is not established and may be because the benzodiazepine eases extrapyramidal adverse effects, allowing an increase in the antipsychotic dose. Benzodiazepines do not influence delirium symptoms, they contain behaviour by inducing sleep.
Short acting benzodiazepines are the drugs of choice - midazolam is ideal, though clonazepam is more commonly used because of its availability and despite its longer half- life. The organically impaired and the elderly may rarely respond paradoxically and their behaviours become more disinhibited. This is more likely in borderline and primitively disturbed personalities or young children and is generally an indication that the dosage used is insufficient to sedate. There is a risk of respiratory depression if parenteral administration is too rapid. Sedation enhances the risk of complications such as falls, DVTs and decubitus ulceration, thus should only be considered in severe and desperate deliria. There appears to be a dosage, above which benzodiazepines are ineffective sedatives, this dose being possibly greater than 6-8 mgs of clonazepam (or midazolam 240 mgs) parenterally over 24 hours. Presumably the benzodiazepine receptors are saturated. Alternative sedatives are chlormethiazole, opioids (in the opioid naive) and barbiturates although the latter drugs are very difficult to titrate to the clinical state.
Very rarely it is necessary to anaesthetise a delirious patient, to render them insensible. If all other interventions have failed and an investigation is essential, or the person requires behavioural control because of the risk of violence and/or the physiological risks of extreme exertion are potentially fatal, anaesthesia may be necessary. Anaesthetic induction with proprofol requires expertise and airway support.
Death from terminal disease is not always peaceful. Modern oncological practices may prolong life but complicate death because of the accumulated disease burden. In palliative care settings terminal sedation is practiced in about 25% of dying patients with intractable symptoms. The commonest indication is that of irreversible delirium, the symptoms have not been controlled despite the best efforts. The intent of deep sedation is to 'kill the symptoms' and not the patient. It is not physician-assisted euthanasia. The doctrine of double effect is the ethical justification for this practice [Table - 9]. Consent may be obtained from the patient during a lucid period, though often family and staff are required to make the decision. Benzodiazepines are the preferred sedative medication. By using a short-acting benzodiazepine such as midazolam, aided if necessary by the antidote flumazenil, the depth of sedation can be carefully titrated. Levomepromazine can be added if the benzodiazepine is insufficient to achieve a settled clinical state. Deep sedation for terminally ill patients may prolong life, possibly for up to several days., Presumably the relief of the distressing symptoms afforded by the sedation allows greater physiological and psychological comfort, whilst awaiting nature's determination of the time of death. When there is no other means of relieving an unendurable symptom, terminal sedation may be the humane clinical option.
To be "crazy", even temporarily, is a distressing and humiliating complication of significant physical illness. Improved recognition and prompt management of delirium, even in the last days of life, can significantly improve the quality of remaining life. Delirium is an extraordinarily difficult condition in which to evaluate treatment outcomes. Clinicians have long recognised the importance and impact of curative interventions (if possible) and palliative interventions (always) in delirium. By increasing the deliriant threshold with multi-component interventions lucid periods may elongate and symptoms subside.
[Table - 1], [Table - 2], [Table - 3], [Table - 4], [Table - 5], [Table - 6], [Table - 7], [Table - 8], [Table - 9]