Antipsychotics for the Prevention and Treatment of Delirium [Internet]

Rockville (MD): Agency for Healthcare Research and Quality (US); 2019 Sep. Report No.: 19-EHC019-EF.


Objectives: To assess benefits and harms of antipsychotics for the prevention and treatment of delirium in adult patient populations.

Data sources: We searched PubMed®, Embase®, the Cochrane Central Register of Controlled Trials (CENTRAL), Cumulative Index to Nursing and Allied Health Literature (CINAHL®), and PsycINFO® through March 2019. We also hand-searched the reference lists of included articles, relevant reviews, and delirium-specific bibliographic repositories.

Review methods: We included randomized controlled trials (RCTs) of antipsychotics that evaluated benefits or harms, and also observational studies that reported harms. Two reviewers independently screened search results for eligibility, serially abstracted data, and independently assessed the risk of bias of the studies and graded the strength of evidence (SOE) for prespecified critical outcomes: delirium severity, cognitive functioning, length of stay in hospital, inappropriate continuation of antipsychotic drugs, falls, sedation, and caregiver burden/strain.

Results: We identified 14 RCTs and 1 observational study evaluating the use of antipsychotics in prevention of delirium. For the treatment of delirium, we identified 19 RCTs and 25 observational studies. Two RCTs were classified as both a prevention and treatment trial. In trials of the prevention of delirium across all populations, there was no difference in delirium incidence for haloperidol versus placebo (relative risk [RR], 0.94; 95% confidence interval [CI], 0.77 to 1.16). Second-generation antipsychotics, compared with placebo, may decrease delirium incidence in postoperative patients at risk for delirium (RR, 0.36; 95% CI, 0.26 to 0.50). Antipsychotics (both haloperidol and second-generation), compared to placebo, demonstrated no differences for length of stay in hospital (low SOE for second-generation antipsychotics and high SOE for haloperidol). We were unable to draw conclusions regarding the effect of antipsychotics on sedation, falls, and delirium severity (insufficient SOE). We found no studies evaluating cognitive functioning, inappropriate continuation of antipsychotic drugs, or caregiver burden/strain. For treatment of delirium, there was little to no difference in effect of haloperidol and second-generation antipsychotics compared with placebo for length of stay in hospital (moderate SOE) and sedation (low and moderate SOE, respectively) with insufficient or no evidence for cognitive functioning or delirium severity. Also, effects of second-generation antipsychotics were not significantly different compared with haloperidol for delirium severity (moderate SOE), cognitive functioning (low SOE), length of stay in hospital (moderate SOE), and sedation (moderate SOE). We found no studies reporting inappropriate continuation of antipsychotic drugs, falls, or caregiver burden/strain. We did not find statistically significant differences for haloperidol or second-generation antipsychotics in neurological harms, including extrapyramidal side effects and neuroleptic malignant syndrome. However, cardiac harms tended to occur more frequently with antipsychotics, specifically prolongation of QT interval with second-generation antipsychotics.

Conclusions: Haloperidol or second-generation antipsychotics, compared to placebo, used for the prevention or treatment of delirium did not improve length of stay in hospital. We found little or no evidence to determine the effect of antipsychotics on cognitive function, delirium severity, or caregiver burden. Second-generation antipsychotics may decrease delirium incidence in postoperative patients, but this evidence is limited and requires more study. We did not detect neurological harms associated with haloperidol or second-generation antipsychotics used for the prevention or treatment of delirium. However, cardiac effects tended to occur more frequently in those receiving antipsychotics. Future studies should include standardized, clinically meaningful measures of patient distress, subsequent memories of delirium, caregiver burden and distress, inappropriate continuation of antipsychotic therapy, and long-term cognitive and functional outcomes.

Publication types

  • Review

Grant support

Prepared for: Agency for Healthcare Research and Quality, U.S. Department of Health and Human Services, 5600 Fishers Lane, Rockville, MD 20857; www.ahrq.govContract No. 290-2015-00006-I, 290-32008-TPrepared by: Johns Hopkins University Evidence-based Practice Center, Baltimore, MD