Delirium

Delirium is an acute, fluctuating disturbance of attention and awareness, together with an additional change in cognition (memory, orientation, language, visuospatial ability, or perception), that develops over hours to a few days and is not better explained by a pre-existing or evolving neurocognitive disorder.^[1] It results from an interaction between predisposing vulnerability (older age, pre-existing cognitive impairment, sensory impairment, frailty) and one or more precipitating insults (infection, metabolic derangement, drug toxicity or withdrawal, surgery, uncontrolled pain, sleep deprivation) — the more predisposing factors present, the smaller a precipitating insult needed to trigger delirium. Delirium is a medical emergency: its presence itself signals a need for prompt evaluation for a potentially life-threatening cause, not just symptom control. It is usually reversible once the precipitating cause is corrected, but unrecognized or poorly managed delirium is associated with prolonged hospitalization and long-term cognitive decline.^[2]

Core features

An acute change from the patient's baseline, with a course that fluctuates over the day (lucid intervals are common and do not exclude delirium), inattention (difficulty focusing, sustaining, or shifting attention — often the most sensitive bedside clue), and disorganized thinking or an altered level of consciousness. Symptoms are frequently worse in the evening and overnight ("sundowning").^[1]

Predisposing factors (increase vulnerability)

• Older age.
• Pre-existing cognitive impairment or dementia — the single strongest predisposing factor.
• Sensory impairment (vision or hearing).
• Frailty and multiple comorbidities.
• Prior episode of delirium.
• Functional impairment / poor baseline mobility.
• Admission to a postoperative, palliative care, or intensive care setting.
• Polypharmacy, particularly a high anticholinergic or sedative burden.
• Alcohol use disorder or chronic benzodiazepine use.

Precipitating conditions (trigger the episode)

A useful systematic framework is the "I WATCH DEATH" mnemonic for categories of acute, potentially reversible causes:

Delirium vs. dementia vs. depression — the classic diagnostic confusion

Distinguishing delirium from dementia is essential because the treatment implications are entirely different: delirium is an acute emergency requiring urgent cause-finding, while dementia is a chronic process managed on a completely different timeline. Depression is the third common confusable diagnosis, particularly with hypoactive delirium.

A patient with pre-existing dementia can develop superimposed delirium — always ask a collateral source (staff, family) whether the current presentation represents an acute change from the patient's own baseline, since "confused" on its own is not informative.

History

• Timeline and course from the patient and, critically, collateral history from staff or family — the single most useful question is whether this represents an acute change from baseline.
• Complete medication review, including over-the-counter products, recent dose changes, and any recently stopped medication (withdrawal).
• Alcohol and substance use history.
• Review of active medical/surgical issues, recent procedures, and pain control.

Examination and investigations

• Full physical examination looking for an infectious source, dehydration, urinary retention or fecal impaction, and new focal neurologic signs.
• Core laboratory work for essentially all patients: complete blood count (including folate and vitamin B12), electrolytes, glucose, renal and liver function, thyroid panel, and urinalysis.
• CT head and/or EEG are not routine — reserve for patients with a new focal neurologic deficit, head trauma, suspected seizure, or another specific clinical suspicion of CNS pathology, rather than ordering them reflexively in every delirious patient.^[3]

Screening tools

Use a validated bedside tool consistently, rather than relying on staff noticing agitation, since this is what allows hypoactive delirium to be caught. The Confusion Assessment Method (CAM) is the most widely used — it is a diagnostic algorithm built from 4 features, not a summed severity score.^[4]

CAM is positive for delirium (diagnosis met) when Features 1 AND 2 are both present, plus EITHER Feature 3 OR Feature 4. A CAM-ICU version is validated for mechanically ventilated or nonverbal patients.^[5] The 4AT is a fast alternative screening tool (alertness, a 4-item cognitive test, an attention test, and acute change/fluctuation) that can be completed in a few minutes without special training and is endorsed alongside CAM by current guidance.^[6]

• Identify and treat or correct the underlying precipitating cause(s) — this is the actual definitive treatment, not a preliminary step before "real" therapy.
• Maintain the safety of the patient, other patients, and staff by managing agitation with the least restrictive effective approach.
• Prevent delirium-related complications — falls, immobility, aspiration, pressure injury, dehydration, and unnecessary indwelling devices/restraints.
• Support the fastest possible return to the patient's cognitive and functional baseline.
• Minimize cumulative antipsychotic and benzodiazepine exposure — use the lowest effective dose for the shortest duration, and taper/discontinue as soon as delirium resolves.
• Avoid iatrogenic worsening from unaddressed deliriogenic medications, unnecessary restraints, or disrupted sleep-wake cycles.

Multicomponent nonpharmacologic measures — first-line for every patient

• Frequent reorientation — clocks, calendars, and family photos in view, with staff and family reminding the patient of the date, place, and reason for admission at each interaction.
• Ensure the patient has their glasses and hearing aids available and in working order — uncorrected sensory impairment is a major, easily fixed contributor.
• Support the normal sleep-wake cycle: encourage daytime light exposure and activity, minimize nighttime noise/light/unnecessary overnight vitals or medication administration, and avoid daytime napping where possible.
• Mobilize the patient as early and as often as safely possible, rather than defaulting to bed rest.
• Avoid physical restraints — they increase agitation and injury risk and can prolong delirium; use the least-restrictive alternative (1:1 observation, bed/chair alarms, moving the patient closer to the nursing station) instead.
• Communicate the diagnosis to the patient (as appropriate) and caregivers, and actively involve family/caregivers at the bedside — familiar presence is itself therapeutic.
• Maintain hydration, nutrition, and normal bowel/bladder function; treat pain adequately (uncontrolled pain is itself a precipitant).
• Structured multicomponent programs such as the Hospital Elder Life Program (HELP) — targeted screening and interdisciplinary management of predisposing risk factors in older inpatients — reduce delirium incidence and fall rates and are the model this approach is based on.^[8]

Medication review and deprescribing — a core pharmacist role

Minimizing or removing drugs that may be contributing to delirium is one of the highest-yield actions a pharmacist can take, and should happen at the same time as, not after, other management steps. Cross-checking the medication list against the current AGS Beers Criteria for potentially inappropriate medications in older adults is a practical way to systematically catch the drug classes below.^[9]

There is insufficient evidence to support using any pharmacologic agent — including antipsychotics or melatonin — purely to prevent delirium; multicomponent nonpharmacologic prevention (the HELP model above) is the evidence-based approach to prevention.^[10]

Pregnancy

There are no controlled studies of pharmacologic delirium treatment in pregnancy. Use non-pharmacologic measures as first-line management, and reserve pharmacotherapy for situations of significant risk of harm to the patient, fetus, staff, or other patients, with informed consent. Antipsychotic use for psychosis in pregnancy has not been associated with major teratogenic malformations, though prepartum use of a high-potency first-generation agent (e.g., haloperidol) has been associated with neonatal hypertonia.^[2]

Breastfeeding

Antipsychotics pass into breast milk at levels generally considered compatible with breastfeeding, though data are limited. If an antipsychotic is used, monitor the infant for lethargy, drowsiness, and (with longer use) developmental milestones. Haloperidol at 10 mg/day or less has not shown adverse infant effects in limited data; higher doses are not recommended. Of the second-generation agents, olanzapine has the most reassuring lactation data; quetiapine also shows low infant exposure; risperidone is associated with relatively more infant exposure and may not be a preferred choice.

Older adults (the population most affected)

Use lower antipsychotic doses than in younger adults (see Pharmacologic Treatment Options for geriatric-specific dosing). Avoid loxapine and olanzapine where possible — both carry significant anticholinergic burden that can worsen cognitive impairment. Avoid benzodiazepines outside the alcohol/benzodiazepine-withdrawal indication; if a benzodiazepine is required for that indication, prefer lorazepam or oxazepam over diazepam given reduced accumulation risk. Older adults, particularly those with pre-existing dementia, carry the FDA boxed-warning mortality and stroke risk with antipsychotic use — weigh this explicitly before starting therapy, use the lowest effective dose, and discontinue promptly once agitation resolves.^[11]

Dementia

Delirium superimposed on dementia is common and easily missed if the change from the patient's own baseline is not actively assessed — see Assessment & Diagnosis for the delirium-vs-dementia comparison. The antipsychotic mortality/stroke boxed warning applies specifically to patients with dementia-related psychosis and is directly relevant whenever an antipsychotic is being considered for a delirious patient who also has dementia.

Renal and hepatic impairment

Reduce antipsychotic doses in significant renal or hepatic impairment and monitor closely for accumulation and QTc prolongation. In hepatic impairment, lorazepam or oxazepam are preferred over diazepam for the withdrawal-delirium indication, since they are cleared by glucuronidation without active metabolites.

Pediatrics

Delirium also occurs in children, particularly postoperatively and in the ICU, using pediatric-specific screening tools (e.g., the Cornell Assessment of Pediatric Delirium). Management principles mirror the adult approach — non-pharmacologic measures and treating the underlying cause first — but pediatric antipsychotic dosing for agitated delirium is not well established; involve pediatric specialist input before using an antipsychotic in a child, rather than extrapolating adult doses.

Reassess delirium status at least daily (more often if agitation is unstable) using the same validated tool (CAM, CAM-ICU, or 4AT) each time, so trajectory — not just a single snapshot — is tracked.^[4] If an antipsychotic was started, monitor sedation level, extrapyramidal symptoms, and QTc (baseline and periodically, especially with risk factors: older age, female sex, bradycardia, pre-existing heart disease, hypokalemia, hypomagnesemia) at each reassessment, and confirm daily that continued therapy is still needed. If a benzodiazepine was started for withdrawal delirium, monitor sedation and respiratory status closely, particularly with concurrent CNS depressants. Reassess the medication list again at discharge to confirm any deliriogenic drug that was held is restarted only if truly needed, and that any antipsychotic or benzodiazepine started acutely has actually been stopped rather than carried forward by default. Once delirium resolves, consider a formal cognitive and functional assessment — delirium can unmask previously undetected baseline cognitive impairment that deserves its own follow-up.

1. Agitation or confusion persists despite nonpharmacologic measures and an adequate low-dose antipsychotic trial: before increasing the dose, reassess for an unaddressed, new, or missed underlying cause — this is the most common reason apparent "antipsychotic failure" actually occurs.
2. If the cause has genuinely been addressed and antipsychotic effect remains inadequate at a reasonable low dose: consider switching agents (e.g., away from haloperidol or risperidone toward quetiapine if parkinsonism or Lewy body disease is present) rather than repeatedly escalating the dose of the same drug.
3. Severe, unsafe agitation persisting despite optimized nonpharmacologic measures and an antipsychotic trial: escalate to psychiatry or geriatric medicine consultation and consider transfer to a higher level of care for closer monitoring.
4. Delirium not resolving after the identified cause has been treated and a reasonable recovery window has passed: reassess for a coexisting or previously missed diagnosis — unrecognized dementia, a new CNS event, ongoing subclinical infection, or drug accumulation from reduced clearance.
5. Alcohol/benzodiazepine withdrawal delirium not controlled by escalating symptom-triggered benzodiazepine dosing: this defines resistant withdrawal — obtain expert/critical-care guidance, since adjunctive therapy under specialist direction may be required.

• Starting an antipsychotic or benzodiazepine as a routine, reflexive response to any delirious patient rather than reserving pharmacotherapy for severe distress or agitation posing a genuine safety risk.
• Missing hypoactive delirium because the patient is quiet and "cooperative" rather than agitated — screen actively with a validated tool instead of relying on staff noticing disruptive behaviour.
• Diagnosing dementia in a patient with an actual acute-onset, fluctuating presentation (or missing delirium superimposed on known dementia) without asking a collateral source about the patient's true baseline.
• Using a benzodiazepine for a delirium etiology other than alcohol/benzodiazepine withdrawal — this typically worsens confusion rather than helping.
• Reaching for physical restraints as a first-line agitation strategy — this increases agitation and injury risk and can prolong the episode.
• Abruptly stopping a chronic benzodiazepine, opioid, or other sedative in a long-term user instead of tapering — precipitates a withdrawal state that itself causes delirium.
• Continuing an antipsychotic started for acute agitation past discharge without a deliberate decision to do so — always taper and discontinue once delirium has resolved.
• Ordering a CT head or EEG reflexively in every delirious patient without a specific clinical indication (new focal deficit, trauma, suspected seizure) rather than reserving it for suspected CNS pathology.
• Starting or continuing a cholinesterase inhibitor (e.g., donepezil) specifically to treat delirium — current evidence does not support benefit for delirium prevention or treatment, so this is not an appropriate use of the drug in this context.
• Overlooking the medication list as a contributing cause — anticholinergic-burden stacking and opioid burden are among the most common, most fixable precipitants a pharmacist will find.

How to use this table: Doses are adult doses for managing agitation associated with delirium and require patient-specific confirmation, including renal/hepatic adjustment and reassessment daily. Select any drug name to open its complete Pharmasist Drug Database page.

Antipsychotics, first-generation

Reserved for severe distress or agitation posing a genuine safety risk, after nonpharmacologic measures and cause-directed treatment; not routine.

Antipsychotics, second-generation

Comparable efficacy to haloperidol in available evidence, with a possibly more favourable adverse-effect profile.

Benzodiazepines — alcohol or benzodiazepine withdrawal delirium ONLY

Not used for any other delirium etiology; using a benzodiazepine outside this specific indication typically worsens confusion rather than helping.

Supportive / adjunctive therapy

Agents without sufficient evidence for routine use

Included here for completeness — none of the following should be started outside a specialist-directed or ICU-specific context; evidence does not currently support routine use for delirium prevention or treatment.