Haloperidol for the Treatment of Delirium
in ICU Patients

Andersen-Ranberg. NEJM 2022. DOI: 10.1056/NEJMoa2211868

Clinical Question

• In patients in the ICU with delirium, does haloperidol compared with placebo, increase the number of days alive and out of the hospital at 90 days?

Background

• Delirium is very common in patients with critical illness and can affect 30-50% of patients admitted to the ICU
• Haloperidol, a typical antipsychotic, is the most frequent agent used to treat delirium in ICU patients despite minimal evidence of its benefit
• The HOPE-ICU study found that early haloperidol did not prevent delirium
• The MIND-USA study reported that the use of haloperidol or ziprasidone had no benefit on the duration of delirium

Design

• Randomised controlled trial
• Computer generated randomisation
• Variable block sizes stratified by centre and delirium motor subtype (hyperactive or hypoactive)
• Blinding of patients, clinicians, and outcome assessors
• Haloperidol and placebo contained in identical ampoules and labels and were indistinguishable from one another
• Intervention ran until death, ICU discharge or 90 days
• 1000 patients required for 90% power to detect 8% improvement in primary outcome, assuming that haloperidol would lead to 15% reduction in in-hospital death
• Intention to treat analysis included patients that were randomised and received assigned study drug
• Secondary analysis of the primary outcome was adjusted for additional predefined risk factors at baseline

Setting

• Screening performed at 18 general ICUs in Europe, of whom 16 ICUs randomised patients
• Dates of data collection: June 2018 – April 2022

Population

• Inclusion:
  • ≥18 years old
  • Admitted to ICU for acute condition
  • Diagnosed with delirium based on either
    • Confusion Assessment Method for the ICU (CAM-ICU) or
    • Intensive Care Delirium Screening Checklist (ICDSC)
• Exclusion:
  • Contraindications to haloperidol
  • Treatment with antipsychotic – either regularly or in ICU prior to inclusion
  • Delirium assessment not possible
  • Alcohol induced delirium/delirium tremens
  • Pregnancy
• 1000 patients randomised, 13 patients excluded post randomisation
  • Primary outcome date obtained from 963 patients
• Comparing baseline characteristics of haloperidol vs. placebo group
  • Age: 70 vs. 71
  • Female: 35 vs. 33%
  • Risk factor for delirium
    • Alcohol overconsumption: 17 vs. 16%
    • Received benzodiazepines in hospital before randomisation: 33 vs. 29%
    • Received haloperidol in hospital before ICU admission: 7 vs. 7%
  • Medical admission type: 64 vs. 69%
  • Surgical admission type: 37 vs. 32%
  • Use of organ support at randomisation
    • Mechanical ventilation: 64 vs. 63%
    • Vasopressors/inotropes: 54 vs. 49%
  • Hypoactive delirium at randomisation: 55 vs. 54%
  • Hyperactive delirium at randomisation: 45 vs. 46%
  • Median type from ICU admission to randomisation: 3.9 vs. 4.1 days

Intervention

• Haloperidol
  • 2.5mg IV TDS
  • Additional doses allowed – up to total of 20mg/day
  • Patients received a median daily dose of 8.3mg for a median of 3.6 days

Control

• Placebo

Management common to both groups

• Patients screened twice daily for delirium
• Treatment of study drug paused if delirium not present when screened (defined as two consecutive negative screening tests)
  • Intervention could be resumed if subsequent delirium screen positive, or readmitted to ICU and delirium present
• Patients could receive rescue medication (propofol, benzodiazepines, or α2 -agonists) at the discretion of the clinical team for uncontrollable delirium
  • Use of other antipsychotic drugs was not permitted

Outcome

• Primary outcome: Number of days alive and out of the hospital by day 90
• Comparing haloperidol vs place group: no significant difference
  • 35.8 vs. 32.9
  • Adjusted absolute difference 2.9 (95% C.I. -1.2 to 7), P=0.22
• Death at 90 days – significantly lower in haloperidol group
  • 36.3 vs. 43.3%
  • Adjusted absolute difference -6.9 (95% C.I. -13 to -0.6)
  • Data missing for 24 patients
  • Fragility index 4 patients
• Mean length of hospital stay – no significant difference
  • 28.8 vs 26.4 days
  • Adjusted absolute difference: 2.3 (95% C.I. -0.6 to 5.1)
• Secondary outcomes:
• Comparing haloperidol vs. placebo group
  • No significant difference in:
    • Days alive without delirium or coma
      • 57.7 vs. 52.6
      • Adjusted absolute difference 5.1 (95% C.I. -1.2 to 11.3)
    • Serious adverse reaction in ICU
      • 2.2% vs. 1.9%
    • Use of rescue medication
      • Total
        • 57.5% vs. 62.1%
      • Propofol
        • 17.6% vs. 15%
      • a2-agonists
        • 47.7% vs. 52.1%
      • Benzodiazepines
        • 27.3% vs. 32.5%
    • Use of open-label antipsychotic
      • 13.2% vs. 13%
    • Use of restraint during delirium
      • 1.8% vs. 2.1%
    • Study drug stopped due to QTc prolongation
      • 2.4% vs. 1.4%
• Subgroup analyses of the primary outcome, comparing haloperidol vs. placebo groups – no significant difference
  • Hyperactive delirium: 39.3 vs. 34.9 days
  • Hypoactive delirium: 33.0 vs. 31.2 days
  • Age ≥69yr: 32.9 vs. 28.7 days

Authors’ Conclusions

• Haloperidol did not lead to a significantly greater number of days alive and out of the hospital at 90 days than placebo

Strengths

• Randomised controlled trial
• Blinded
• Multi-centre
• Validated assessment tools for diagnosing delirium

Weaknesses

• Low number of patients from international sites may limit external validity
• Composite outcome
• Power calculation based on reduction in mortality of 15% which may have been overly optimistic
• Use of open label antipsychotics was 13%, although similar in both groups
• High rate of major protocol violations ~20%
• 25% of those screened were excluded due to prior antipsychotic use in the ICU. This may lead to a selection bias
• High rates of hypoactive delirium treated with haloperidol which is not standard practice in my institution

The Bottom Line

• In patients admitted to the ICU who develop delirium, the use of haloperidol compared with placebo did not significantly improve the number of days alive and out of the hospital by day 90
• The mortality benefit noted with the use of haloperidol needs to be treated with caution due to the fact that it was not the primary outcome, and the number of patients with missing data exceeds the fragility index
• In patients with hyperactive delirium I will continue to use haloperidol to treat symptoms despite the minimal evidence for its benefit on longer term outcomes. Larger trials would be needed to investigate any mortality benefit

External Links

• article Haloperidol for the Treatment of Delirium in ICU Patients
• further reading Management of Delirium in the ICU
• further reading Medicines & Healthcare products Regulatory Agency: Review of the use of haloperidol in elderly patients with acute delirium

Metadata

Summary author: @davidslessor
Summary date: 1st December 2022
Peer-review editor: George Walker

Picture by: Pixabay/M Wewering