From Wiki Journal Club
Jump to navigation Jump to search
Andersen-Ranberg NC, et al. "Haloperidol for the Treatment of Delirium in ICU Patients". The New England Journal of Medicine. 2022. Online ahead of print(2022-10-26):1-11.

Clinical Question

In adult patients admitted to the ICU that develop hyperactive or hypoactive delirium, does scheduled and as needed intravenous haloperidol as compared to placebo affect the number of days alive and out of the hospital?

Bottom Line

Among ICU patients with hyperactive or hypoactive delirium, scheduled+PRN haloperidol did not lead to greater number of days alive and out of the hospital at 90 days. Haloperidol was, however, associated with lower all-cause mortality.

Major Points

Hyperactive and hypoactive delirium occur frequently among critically ill patients, with these conditions affecting up to 50% of this population. Delirium is have higher burden of morbidity and mortality. The 2018 MIND-USA trial randomized 566 ICU patients with delirium to haloperidol, ziprasidone, or placebo, that trial found no statistical difference in their primary outcome of days alive without delirium or coma, there was also no difference in all-cause mortality, though haloperidol nearly met statistical significance for worsening mortality at 90d versus placebo (haloperidol 38% vs. placebo 34%, OR 1.17; 95% CI 0.99-1.40).

Published in 2022, the Agents Intervening against Delirium in the Intensive Care Unit (AID-ICU) trial randomized 1,000 patients with ICU delirium to haloperidol (2.5mg IV TID + PRN to maximum of 20 mg/d) or placebo. Specifically, these were adult patients admitted to an ICU for an acute cause that screened positive for delirium with either the CAM-ICU or ICDSC. They found no difference in their primary outcome of Days alive and out of hospital at 90 days (haloperidol 35.8 days vs. placebo 32.9 days, P=0.22). All-cause mortality at 90 days was lower in the haloperidol group (36.4% vs. 43.3%; adjusted RR 0.84; 95% CI 0.72 to 0.98). Adverse events were similar between groups.

Taken together with MIND-USA, AID-ICU provides evidence that scheduled+PRN haloperidol among ICU patients with delirium may not alter days alive and out of the hospital, but might lead to a mortality benefit. In general, haloperidol is not associated with more adverse events in delirium.


Clinical Practice Guidelines for the Prevention and Management of Pain, Agitation/Sedation, Delirium, Immobility, and Sleep Disruption in Adult Patients in the ICU (2018, adapted)[1]

  • Prevention: suggest not using haloperidol, dexmedetomidine, or statin
  • Treatment: suggest not using haloperidol or a statin
  • Note: These guidelines were published prior to the publication of AID-ICU


  • Multicenter, double-blind, randomized, placebo-controlled trial
  • N=1000
    • Haloperidol (n=510)
    • Placebo (n=490)
  • Setting: 16 general ICUs in Denmark, Finland, the United Kingdom, Italy, and Spain
  • Enrollment: 2018-2022
  • Follow-up: 90 days
  • Analysis: Per-protocol
  • Primary Outcome: number of days alive and out of the hospital at 90 days after randomization


Inclusion Criteria

  • ≥18 years old with unplanned admission to an ICU for an acute condition
  • Screened positive for delirium with 1 of the following tools:
    • Positive result on the Confusion Assessment Method for the ICU (CAM-ICU)
    • Score ≥4 on the Intensive Care Delirium Screening Checklist (ICDSC)

Exclusion Criteria

  • Haloperidol contraindication (intolerance to drug, known Parkinson’s disease, known QTc prolongation, history of tardive dyskinesia)
  • Non-pharmacologic coma
  • Regular anti-psychotic therapy
  • Permanently incompetent patient
  • Withdrawal from active therapy
  • Brain death
  • Pregnancy
  • Involuntary hospitalization
  • Alcohol withdrawal

Baseline Characteristics

Haloperidol Group displayed

  • Demographics: median age 70 years, 35% female
  • Delirium risk-factors: 1.6^ TBI, 2.4% previous stroke, 5.8% mental illness, 0.4% neurodegenerative disease, 31% smoker, 17% over-consumption of ethanol, 1.8% substance abuse, 2.6% pre-hospital exposure to benzodiazepines, 33% pre-randomization exposure to benzodiazepines
  • Admission type: 36% surgical, 63% medical, median time from ICU admission to randomization 3.9 days
  • Organ support: 64% mechanically ventilated, 54% vasopressor/ionotropes, 15% renal replacement therapy
  • Delirium subtype: 55% hypoactive, 45% hyperactive


  • Randomization to a group:
    • Haloperidol at 2.5mg IV three times daily, plus as needed doses of 2.5mg up to a daily maximum of 20mg
    • Matched placebo of both scheduled and as needed doses
  • Rescue medications (benzodiazepines, propofol, or alpha2-agonists) could be used at clinician discretion
  • Patients who were comatose continued to receive trial medications.
  • All patients had their delirum assessed 2x/day with CAM-ICU or ICDSC tools; treatment paused if no longer delirious on 2 consecutive CAM-ICU or ICDSC tools on the same day, though was resumed if these again became positive
  • Treatment stopped upon ICU discharge, at 90 days, or death


Comparisons are haloperidol vs. placebo.

Primary Outcomes

Days alive and out of hospital at 90 days
35.8 vs. 32.9 (adjusted difference 2.9; 95% CI -1.2 to 7; P = 0.22)

Secondary Outcomes

All-cause mortality at 90 days
Table 3 notes that this metric was missing for 1 patient in the placebo group, and also missing in 24 patients overall. The WJC editorial staff are unclear of the missingness, but suspect the note of 24 patients missing this outcome is a typo. Also of note, the authors present this in Table 3 as a primary outcome but the listing notes this is a secondary outcome.
36.3% vs. 43.3% (adjusted difference -6.9; 95% CI -13 to -0.6; Adjusted RR 0.84; 95% CI 0.72 to 0.98)
Length of hospital stay
The authors present this in Table 3 as a primary outcome but the listing does not identify this as a primary our secondary outcome.
28.8 vs. 24.4 days (adjusted difference 2.3; 95% CI -0.6 to 5.1)
Days alive
...without delirium or coma: 57.7 vs 52.6 days (adjusted difference 5.1; 99% CI -1.2 to 11.3)
...without mechanical ventilation: 57.9 vs. 53.9 days (adjusted difference 4; 99% CI -2.2 to 10.1)

Adverse Events

Serious adverse reaction in ICU (a secondary outcome)
2.2% vs. 1.9% (adjusted difference 0.4; 99% CI -1.9 to 2.7)
Use of rescue medication (a secondary outcome)
57.5% vs. 62.1% (adjusted difference -4; 99% CI -11.8 to 3.6)
Days with use of rescue medication per patient (a secondary outcome)
2.9 vs. 2.9 days (adjusted difference 0.1; 99% CI -0.7 to 0.9)
Therapy stopped due to QTc prolongation
2.4% vs. 1.4%
Physical restraints utilized
1.9% vs. 2.1%

Subgroup Analysis

There was no obvious heterogeneity for the primary outcome by type of delirium, age, sex, admission type, whether risk factors for delirium were present, and SMS-ICU score.


  • Unclear generalizability outside of the included countries
  • Might have underrecruited individuals with hypoactive delirium as individuals with this condition are less clinically apparent
  • Underreporting of medical comorbidities, use of other pharmacological medications
  • May not have been correctly powered given the composite endpoints
  • All-cause mortality did not use time-to-event analyses with Cox proportional hazards modeling


Presumably funding is from public funders and non-profits.

Further Reading