The primary treatment of the symptoms of delirium is largely pharmacologic. The high-potency antipsychotic medication haloperidol is most frequently employed, although other pharmacologic and somatic interventions have been used in particular instances. Recently, there has been increased use of risperidone (61, 62). The available studies of the efficacy and other outcomes from use of these treatments for patients with delirium are reviewed in this section.

Several important points should be considered when evaluating the evidence for specific somatic interventions. While haloperidol has been the most studied pharmacologic treatment, few studies have used a standardized definition of delirium (e.g., based on DSM-IV criteria). In addition, few investigations have used reliable and valid delirium symptom rating measures to assess symptom severity before and after intervention.

For somatic treatments other than haloperidol, there have been no large, prospective trials or studies including a control group. Information regarding the efficacy of these treatments comes mainly from small case series or case reports; interpretation of the results from many of these case presentations is also made difficult by the use of nonstandardized definitions of delirium or informal measures of delirium symptom severity.


1. Antipsychotics

a) Goals and efficacy

Antipsychotics have been the medication of choice in the treatment of delirium. Evidence for their efficacy has come from numerous case reports and uncontrolled trials (63, 64). A series of controlled trials also showed that antipsychotic medications can be used to treat agitation and psychotic symptoms in medically ill and geriatric patient populations (65–69). However, most of these trials were not conducted with patients who had clearly or consistently defined delirium; in some studies, agitation and disorientation were the sole criteria and symptom assessments ranged from questionnaires to simple identification without symptom descriptions.

A randomized, double-blind, comparison trial by Breitbart et al. (70) identified delirium by using standardized clinical measures, and it demonstrated the clinical superiority of antipsychotic medications over benzodiazepines in delirium treatment. The Delirium Rating Scale, Mini-Mental State examination, and DSM-III-R were used to make the diagnosis in 244 hospitalized AIDS patients. The subjects were randomly assigned to one of three medications: chlorpromazine, haloperidol, and lorazepam. There were statistically significant decreases in scores on the Delirium Rating Scale after 2 days in the haloperidol and chlorpromazine groups but not in the lorazepam group (the mean decreases in scores were 8.0, 8.5, and 1.0, respectively). The improvement in delirium symptoms observed among those treated with antipsychotic medications occurred quickly, usually before the initiation of interventions directed at the medical etiologies of the delirium.

Droperidol, a butyrophenone with a rapid onset of action and relatively short half-life that is more sedating than haloperidol, has also been found to be an effective treatment for hospitalized patients with agitation, although not necessarily delirium (71). Results of two double-blind clinical trials comparing droperidol to haloperidol suggest that a more rapid response may be obtained with droperidol. Resnick and Burton (72) reported that 30 minutes after intramuscular injections, 81% of patients initially treated with 5 mg of haloperidol required a second injection, compared to only 36% of patients initially given 5 mg of droperidol. Thomas and colleagues (69), comparing 5 mg i.m. of droperidol to 5 mg i.m. of haloperidol, found significantly decreased combativeness among the droperidol treatment group after 10, 15, and 30 minutes. There has been very little study of the newer antipsychotic medications (risperidone, olanzapine, and quetiapine) in the treatment of delirium. Although there have been several case reports of use of risperidone for patients with delirium (61, 62, 73, 74), there have been no published clinical trials of any of the new antipsychotic medications for patients with delirium.

b) Side effects

Phenothiazines can be associated with sedation, anticholinergic effects, and α‐adrenergic blocking effects that can cause hypotension; each of these side effects may complicate delirium. Butyrophenones, particularly haloperidol and droperidol, are considered the safest and most effective antipsychotics for delirium. Haloperidol, a high-potency dopamine-blocking agent with few or no anticholinergic side effects, minimal cardiovascular side effects, and no active metabolites, has generally been considered the antipsychotic medication of first choice in the treatment of delirium. High-potency antipsychotic medications also cause less sedation than the phenothiazines and therefore are less likely to exacerbate delirium. Although droperidol may have the advantages of a more rapid onset of action and a shorter half-life than haloperidol, droperidol is associated with greater sedation and hypotensive effects (75).

The use of antipsychotic medications can be associated with neurological side effects, including the development of extrapyramidal side effects, tardive dyskinesia, and neuroleptic malignant syndrome. However, there is some evidence to suggest that extrapyramidal side effects may be less severe when antipsychotic medications are administered intravenously (76). One case series involved 10 consecutive general medical inpatients receiving doses of oral or intravenous haloperidol at approximately 10 mg/day. Four patients were given intravenous medication, and six were given oral doses. Although delirium was not identified as the reason for treatment, five patients met diagnostic criteria by description. There was no significant difference in the incidence of akathisia, but the group receiving intravenous medication experienced less severe extrapyramidal symptoms. Neither method of administration resulted in acute dystonic reactions or changes in blood pressure or pulse rate (76).

Haloperidol used in the treatment of delirium has been found in some instances to lengthen the QT interval, which can lead to torsades de pointes, a form of polymorphic ventricular tachycardia that can degenerate to ventricular fibrillation and sudden death. Estimates of the incidence of torsades de pointes among patients with delirium treated with intravenous haloperidol have ranged from four out of 1,100 patients (77) to eight out of 223 patients (78). Although development of this serious event has been associated with higher intravenous doses (>35 mg/day) of haloperidol, it is important to note that torsades de pointes has also been reported with low-dose intravenous haloperidol and oral haloperidol as well (78, 79). Droperidol has also been associated with lengthening of the QT interval, and it may also be associated with torsades de pointes and sudden death.

Other side effects of antipsychotic medication use can rarely include lowering of the seizure threshold, galactorrhea, elevations in liver enzyme levels, inhibition of leukopoiesis, neuroleptic malignant syndrome, and withdrawal movement disorders.

c) Implementation

Although different antipsychotic medications can be given orally, intramuscularly, or intravenously, in emergency situations or when there is lack of oral access, intravenous administration may be most effective. In addition, as described in the preceding section on side effects, there is some evidence that antipsychotic medications may cause less severe extrapyramidal side effects when administered intravenously (76). Intravenous administration of haloperidol has not yet received approval by the Food and Drug Administration (FDA).

There have been few studies to determine the optimal doses of antipsychotic medications in the treatment of delirium. On the basis of doses used in several studies, starting haloperidol in the range of 1–2 mg every 2–4 hours as needed has been suggested (80). Low doses, for example as low as 0.25–0.50 mg of haloperidol every 4 hours as needed, have been suggested for elderly patients (81). On the other hand, severely agitated patients may require titration to higher doses. Bolus intravenous haloperidol doses exceeding 50 mg with total daily doses up to 500 mg have been reported, and they were associated with minimal effects on heart rate, respiratory rate, blood pressure, and pulmonary artery pressure and minimal extrapyramidal side effects (82, 83).

Several studies (75, 84) have examined the use of continuous intravenous infusions of haloperidol or droperidol among agitated medically ill patients who have required multiple bolus intravenous injections of antipsychotic medications. The results indicate that this means of administration can be safe and may help avoid some of the complications associated with repeated bolus dosing (e.g., hypotension). The authors of one study (84) recommended continuous infusion of haloperidol for patients who required more than eight 10-mg haloperidol boluses in 24 hours or more than 10 mg/hour for more than 5 consecutive hours. They suggested initiating haloperidol with a bolus dose of 10 mg followed by continuous infusion at 5–10 mg/hour.

Because antipsychotic medications used in the treatment of delirium have occasionally been found to lengthen the QT interval, possibly leading to torsades de pointes, ventricular fibrillation, and sudden death, recommendations for medication management include a baseline ECG with special attention paid to the length of the QTc interval. A prolongation of the QTc interval to greater than 450 msec or to greater than 25% over that in previous ECGs may warrant telemetry, a cardiology consultation, and dose reduction or discontinuation (85, 86). It has also been recommended that serum levels of magnesium and potassium be monitored in critically ill patients, especially those whose baseline QTc interval is 440 msec or longer, those who are receiving other drugs that increase the QT interval, or those who have electrolyte disturbances (87).


2. Benzodiazepines

a) Goals and efficacy

Few controlled studies have evaluated the efficacy of benzodiazepines as a monotherapy (i.e., not in combination with other pharmacotherapies) for the treatment of delirium. The limited data that are available suggest that benzodiazepine monotherapy may be ineffective as a treatment for general cases of delirium caused by a variety of etiologies. For example, the comparison by Breitbart et al. (70), described in Section III.C.1, indicated that lorazepam, given alone, was less effective as a treatment for delirium than either haloperidol or chlorpromazine.

Although there appears to be little evidence to support the use of benzodiazepines alone for general cases of delirium, there may be certain types of delirium for which benzodiazepines have advantages and are preferable. For example, benzodiazepines are the treatment of choice for delirium related to alcohol or benzodiazepine withdrawal. Other specific clinical circumstances in which benzodiazepines may be useful include instances when there is a need for a medication that can raise the seizure threshold (unlike antipsychotics, which lower the seizure threshold) or when anticholinergic side effects or akathisia associated with antipsychotics would seriously exacerbate a patient's condition.

There have been several reports of the combination of antipsychotics and benzodiazepines for the treatment of delirium, and the results indicate that this combination may decrease medication side effects and potentially increase clinical effectiveness in special populations, for example severely ill cancer patients or AIDS patients. Results of several open studies using intravenous haloperidol along with intravenous lorazepam suggest that the combined treatment is more efficacious, with a shorter duration of the delirium and fewer extrapyramidal symptoms, than intravenous haloperidol alone (16, 63, 88). Most of these studies defined delirium according to DSM criteria but did not use standardized assessment tools.

b) Side effects

The adverse effects of benzodiazepines on mental status have received some attention. Marcantonio et al. (89) demonstrated an association between benzodiazepine use and postoperative delirium in a prospective study of psychoactive medications given to patients admitted for elective noncardiac procedures. Long-acting benzodiazepines in particular posed problems. Benzodiazepines have been associated with sedation, behavioral disinhibition, amnesia, ataxia, respiratory depression, physical dependence, rebound insomnia, withdrawal reactions, and delirium. Geriatric populations are at greater risk for the development of these complications; children and adolescents may also be at increased risk for disinhibition reactions, emotional lability, increased anxiety, hallucinations, aggression, insomnia, euphoria, and incoordination (16, 90, 91).

Benzodiazepines are generally contraindicated in delirium from hepatic encephalopathy due to accumulation of glutamine, which is related chemically to γ-aminobutyric acid (GABA). Benzodiazepines should also be avoided, or used with caution, in patients with respiratory insufficiency. For patients who have hepatic insufficiency or are taking other medications metabolized by the cytochrome P450 system, benzodiazepines that are predominantly metabolized by glucuronidation (lorazepam, oxazepam, and temazepam) should be used when a benzodiazepine is required.

c) Implementation

When benzodiazepines are used, relatively short-acting medications with no active metabolites (e.g., lorazepam) should be selected.

Few studies have investigated the optimal dose of benzodiazepines for the treatment of delirium. However, the dose must be carefully considered, given the possibility that benzodiazepines may exacerbate symptoms of delirium. In cases of delirium due specifically to alcohol or sedative-hypnotic withdrawal, higher doses of benzodiazepines and benzodiazepines with longer half-lives may be required.

In a report of a case series of 20 critically ill cancer patients for which benzodiazepines and antipsychotics were administered together, Adams et al. (63) suggested that treatment be started with 3 mg i.v. of haloperidol followed immediately by 0.5–1.0 mg i.v. of lorazepam. Additional doses and the frequency are then titrated to the patient's degree of improvement. For example, Adams et al. stated that if little or no improvement is observed within 20 minutes, an additional dose of 5 mg i.v. of haloperidol and 0.5–2.0 mg i.v. of lorazepam can be given. In some cases of severe agitation, the eventual doses of both medications have been quite large (e.g., daily doses of lorazepam between 20 and 30 mg and of haloperidol between 100 and 150 mg).


3. Cholinergics

a) Goals and efficacy

Anticholinergic mechanisms have been implicated in the pathogenesis of many medication-induced deliriums. In addition, anticholinergic mechanisms may be involved in delirium from hypoxia, hypoglycemia, thiamine deficiency, traumatic brain injury, and stroke (49). However, cholinergic medications have been used in a very limited fashion to treat delirium, almost exclusively in cases of delirium clearly caused by anticholinergic medications. Physostigmine, a centrally active cholinesterase inhibitor, has been used most often, with tacrine and donepezil receiving less attention.

In one prospective study (92), physostigmine reversed delirium among 30 patients in a postanesthesia recovery room, in whom either atropine or scopolamine had caused the delirium. In four single case reports of delirium diagnosed by clinical interviews, physostigmine reversed the delirium resulting from ranitidine (93), homatropine eyedrops (94), benztropine (95), and meperidine (96).

In a single case study (97), tacrine reversed delirium induced by anticholinergic medication. Newer cholinesterase inhibitors with fewer side effects than tacrine have not been studied for treatment of delirium.

b) Side effects

Many side effects of cholinesterase inhibitors are caused by cholinergic excess; such effects include bradycardia, nausea, vomiting, salivation, and increased gastrointestinal acid. Physostigmine can cause seizures, particularly if intravenous administration is too rapid (98). Tacrine has been associated with asymptomatic increases in liver enzyme levels. A threefold increase has been observed in approximately 30% of patients and is generally reversible with discontinuation of treatment; 5%–10% of patients develop more marked (e.g., tenfold) but still generally reversible increases in liver enzyme levels that warrant discontinuation of tacrine treatment (99).

c) Implementation

Physostigmine is usually administered parenterally. Doses that have been used in studies of delirium have included intravenous or intramuscular injections ranging from 0.16 to 2.00 mg and continuous intravenous infusions of 3 mg/hour (92–96).

In the single case study of tacrine used to reverse delirium induced by anticholinergic medication, 30 mg i.v. was used (97).


4. Vitamins

Certain vitamin deficiencies are commonly described as causing delirium. Consequently, one would expect such deliria to reverse at least to some extent with repletion of the deficient vitamin. Although this has not been subjected to rigorous trials, there are some case reports and case series supporting this effect. A malnourished hemodialysis patient with nicotinamide deficiency had a paranoid delirium that responded to parenteral nicotinamide, 500 mg/day (100). Bahr et al. (101) reported that of two chronic alcoholic patients with B vitamin deficiency, malnutrition, and central pontine myelinolysis, one improved quickly with intravenous vitamins. Thiamine deficiency delirium (DSM-III-R) was treated with vitamin B complex in six of 13 elderly medically ill patients, but only one patient had a dramatic response to treatment (102).

In one randomized controlled trial (103), 26 elderly patients undergoing orthopedic surgery received treatment with intravenous vitamins B and C preoperatively and postoperatively and were compared to 32 age-matched surgical control subjects who did not receive vitamins. There was no difference between the intervention and control groups in the incidence of postoperative confusion (39% versus 38%) or in the preoperative thiamine status as determined by serum assays.

In general, any patient with delirium who has a reason to be B vitamin deficient (e.g., alcoholic or malnourished) should be given multivitamin replacement.


5. Morphine and paralysis

Hypoxia, fatigue, and the metabolic consequences of overexertion all exacerbate delirium. Such hypercatabolic conditions are likely to accompany certain causes of agitated delirium (e.g., hyperdynamic heart failure, adult respiratory distress syndrome, hyperthyroid storm). For such patients and for any cases of agitated delirium unresponsive to other pharmacologic interventions, the patient may require a paralytic agent and mechanical ventilation. This improves oxygenation and reduces skeletal muscle exertion. The patient is usually heavily sedated. Morphine (or other opiate) is also an important palliative treatment in cases of delirium where pain is an aggravating factor (104). However, some opiates can exacerbate delirium, particularly through their metabolites, which possess anticholinergic activity (89). Among opiates, meperidine and fentanyl are particularly anticholinergic.


6. ECT

ECT has not been shown to be an effective treatment for general cases of delirium. Earlier case reports and case series had significant limitations: standardized diagnostic criteria and rating scales were not used; patients with schizophrenia, mania, postpartum psychosis, or psychotic depression were included and diagnosed with delirium because they had disorganized thinking and cognition; and few details concerning the manner in which ECT was performed were provided (105–111).

There is limited evidence for ECT as a treatment for particular cases of delirium due to specific etiologies. MEDLINE literature searches identified two case reports of ECT use for the delirium that is a component of neuroleptic malignant syndrome. In one study (112), the delirium symptoms improved in 24 of 29 patients with neuroleptic malignant syndrome who were treated with ECT. In the second study (113), 26 of 31 patients with neuroleptic malignant syndrome who were treated with ECT and hydration were described as having improved delirium symptoms. ECT has been studied in small samples of patients with delirium tremens. In one older study (109), 10 patients receiving ECT and conventional treatment had a shorter duration of delirium symptoms than 10 patients receiving conventional treatment alone (mean, 0.85 versus 2.8 days, respectively). In one case report (106), a patient with delirium tremens who had not responded to high-dose benzodiazepine treatment was described as recovering after ECT. In two case reports (114, 115), patients with protracted courses of delirium after traumatic brain injuries improved after receiving ECT. Because of the lack of compelling evidence, as well as the availability of alternative means of management, ECT is not presently used in the United States for treatment of delirium tremens.

In addition to the very limited evidence that ECT is an effective treatment for delirium, there may be considerable risks with ECT in medically unstable patients. For these reasons, ECT should be considered only rarely for patients with delirium due to specific etiologies such as neuroleptic malignant syndrome and should not be considered initially as a substitute for more conservative and conventional treatments. ECT itself may carry a risk of both postictal delirium (lasting minutes to hours) and interictal delirium (lasting days) after the procedure (116–120). Beyond that time period, ECT can also exacerbate cognitive deficits, such as memory impairment. Certain patient populations at higher risk for these adverse effects from ECT include patients with Parkinson's disease (particularly those taking carbidopa), Huntington's disease, or caudate or other basal ganglia strokes (119, 121–125).


Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).
Related Content
Manual of Clinical Psychopharmacology, 7th Edition > Chapter 4.  >
The American Psychiatric Publishing Textbook of Psychiatry, 5th Edition > Chapter 8.  >
The American Psychiatric Publishing Textbook of Psychiatry, 5th Edition > Chapter 26.  >
The American Psychiatric Publishing Textbook of Psychiatry, 5th Edition > Chapter 37.  >
Psychiatric News
  • Print
  • PDF
  • E-mail
  • Chapter Alerts
  • Get Citation