Review: Oh ES et al, Ann Intern Med 2019;171(7):474–484; Nikooie R et al, Ann Intern Med 2019;171(7):485–495

Study type: Systematic reviews of randomized controlled trials and prospective studies

Delirium is common and costly. With an incidence rate approaching 50% in hospitalized older adults, the condition poses a burden that is anticipated to grow as the largest elderly population in human history enjoys increasing life expectancy. For patients, delirium heightens the risk of physical and cognitive decline and overall mortality.

Antipsychotics remain a mainstay of delirium treatment despite significant concerns about their safety and efficacy. A pair of reports recently evaluated the evidence on their use in delirium. The first investigated antipsychotics for the prevention of delirium in high-risk (vulnerable) patients. The other report looked at whether antipsychotics can improve symptoms of delirium after it develops.

The reports employed a rigorous protocol developed with input from a panel of experts. They included randomized controlled trials as well as prospective observational studies with comparison groups. All studies used standard diagnostic instruments, and five outcomes were selected for data extraction: cognitive functioning, hospital length of stay, delirium severity, sedation, and inappropriate continuation of antipsychotics. Other outcomes included delirium incidence and duration, mortality, and cardiac and neurological adverse effects.

The prevention analysis included 14 randomized controlled trials, totaling 4281 patients. Most of the studies involved post-surgical interventions. The majority used haloperidol (10 studies), while the other four studies used olanzapine, risperidone, and ziprasidone. No differences were found between haloperidol and placebo in delirium incidence, duration, hospital length of stay, and mortality. Little to no evidence supported haloperidol over placebo in cognitive functioning, delirium severity, sedation, and inappropriate continuation. Second-generation antipsychotics, but not haloperidol, showed a slightly reduced incidence of delirium. However, this difference only proved significant in a pooled analysis of three studies, indicating a weak association.

The treatment analysis included 16 randomized controlled trials and 10 prospective studies (a total of 5607 adult inpatients with delirium). No differences were found between haloperidol and second-generation antipsychotics vs placebo for sedation, delirium duration, hospital length of stay, or mortality. More importantly, there was insufficient evidence to support any antipsychotic over placebo in delirium severity and cognitive functioning. For second-generation antipsychotics, evidence was generally lacking for most outcomes, but there were no differences in overall mortality.

Neither report found evidence of neurological harms associated with short-term antipsychotic use, such as extrapyramidal side effects. However, a higher frequency of cardiac adverse effects (eg, QT interval prolongation) was noted. Both studies had similar limitations, including high variability in patient characteristics, treatment context, antipsychotic dosing, and outcome reporting.

TCPR’s Take
Antipsychotics are not a panacea for all agitated or psychotic states, and their expansion into delirium is not supported by the evidence. Patients at risk for delirium are also more likely to develop problems on these medications, such as QT prolongation, hypothermia, hypotension, and anticholinergic effects like urine retention, constipation, and (frankly) delirium.