sedation in the ICU
Contents
Analgesia first, sedation second. Target light sedation (RASS 0 to −2) in most ventilated patients — deep sedation prolongs ventilation, increases delirium, and worsens mortality. Avoid benzodiazepines (↑delirium risk). Propofol or dexmedetomidine are preferred agents. Pair daily sedation assessment with spontaneous awakening + breathing trials (SAT/SBT). Screen for delirium twice daily with CAM-ICU.
analgesia first
Untreated pain drives agitation more than anxiety does. Assess pain before adding sedation.
- Assessment: Numeric Rating Scale (NRS) if communicating; Behavioural Pain Scale (BPS) or Critical-Care Pain Observation Tool (CPOT) if not
- First-line: intermittent IV fentanyl 25–50 mcg q1–2h or infusion 25–100 mcg/h; hydromorphone 0.2–0.5 mg IV q2–4h
- Adjuncts: acetaminophen 1 g IV/PO q6h (ceiling 4 g/day; 2 g/day if hepatic impairment); ketamine low-dose infusion (0.1–0.3 mg/kg/h) for opioid-sparing or refractory pain
- Avoid meperidine (normeperidine seizures, especially in renal impairment)
sedation targets and scales
| RASS score | term | description |
|---|---|---|
| +4 | combative | violent, danger to staff |
| +3 | very agitated | pulls tubes/lines, aggressive |
| +2 | agitated | frequent non-purposeful movement |
| +1 | restless | anxious, not aggressive |
| 0 | alert and calm | |
| −1 | drowsy | sustained awakening to voice (>10 s) |
| −2 | light sedation | briefly awakens to voice (<10 s) |
| −3 | moderate sedation | movement or eye opening to voice, no eye contact |
| −4 | deep sedation | no response to voice, movement to physical stimulation |
| −5 | unarousable | no response to voice or physical stimulation |
- Target RASS 0 to −2 for most patients
- Deep sedation (RASS −4 to −5) only for: neuromuscular blockade, severe ARDS with proning/high driving pressures, status epilepticus, raised ICP, open abdomen
- Reassess target at least q12h — step down as soon as the indication resolves
sedative agents
| agent | onset | offset | pros | cons |
|---|---|---|---|---|
| propofol | seconds | minutes | rapid titration, ↓ICP, antiemetic | hypotension (dose-dependent), hypertriglyceridaemia, propofol infusion syndrome (PRIS) at >80 mcg/kg/min or >48 h |
| dexmedetomidine | 15–30 min | 1–2 h | cooperative sedation (arousable), no respiratory depression, ↓delirium incidence | bradycardia, hypotension, ceiling effect (~1.5 mcg/kg/h), cost, loading dose often causes hypotension → frequently omitted |
| midazolam | 2–3 min | prolonged (accumulates) | seizure control, anxiolysis | ↑delirium (independent risk factor), unpredictable offset especially with renal/hepatic impairment, tolerance |
| ketamine | 1–2 min | 15–30 min | bronchodilation, haemodynamic stability, opioid-sparing | emergence reactions (attenuated at sub-anaesthetic doses), ↑secretions, ↑ICP (disputed at low doses) |
Rare but lethal. Risk factors: >80 mcg/kg/min, >48 h, catecholamine/steroid co-administration, young adults. Features: unexplained metabolic acidosis, rhabdomyolysis, hyperkalaemia, renal failure, cardiac dysfunction, lipaemic serum. Stop propofol immediately — no specific treatment. Check CK and triglycerides daily when running high-dose propofol.
Propofol for rapid titration needs (e.g. neurosurgical, short-term post-op). Dexmedetomidine when arousability matters (extubation pathway, non-intubated NIV, delirium-prone patients). In practice, many units use propofol for the first 24–48 h then transition to dexmedetomidine as extubation approaches.
daily sedation protocol
SAT + SBT bundle — paired awakening and breathing trials reduce ventilator days and mortality.
- Safety screen — no active seizures, no alcohol withdrawal requiring escalating doses, no neuromuscular blockade, FiO2 ≤0.5 and PEEP ≤8 (institutional thresholds vary)
- SAT — stop or halve sedative infusions; observe for 30 min
- SBT — if SAT passed, trial pressure support (5–8 cmH2O) or T-piece for 30–120 min
- Fail criteria — agitation (RASS ≥ +2), SpO2 <88%, RR >35, new arrhythmia → restart sedation at half the prior dose
Kress (NEJM 2000) — daily interruption reduced ventilator days and ICU LOS. ABC Trial (Lancet 2008) — pairing SAT + SBT improved survival vs SBT alone. SLEAP (JAMA 2012) — adding daily interruption to a protocolised light-sedation strategy showed no added benefit → the protocol itself matters, not necessarily the interruption.
delirium
Affects 50–80% of ventilated ICU patients. Independent predictor of ↑mortality, ↑ventilator days, long-term cognitive impairment.
screening
- CAM-ICU — validated, binary (positive/negative), performed at RASS ≥ −3
- Screen at least twice daily (each nursing shift)
- Hypoactive delirium (RASS −1 to −3, inattentive) is underdiagnosed and carries similar prognosis to hyperactive
prevention (ABCDEF bundle)
- Assess/treat pain
- Both SAT + SBT daily
- Choice of sedative — avoid benzodiazepines
- Delirium screening
- Early mobilisation
- Family engagement
treatment
- Non-pharmacological first: reorientation, day-night cycling (lights, noise reduction), mobilisation, remove unnecessary lines/restraints, correct sensory deficits (glasses, hearing aids)
- Pharmacological: no agent has proven mortality benefit
- MIND-USA (NEJM 2018) — haloperidol and ziprasidone vs placebo → no difference in delirium-free days
- Antipsychotics may reduce distressing symptoms in hyperactive delirium; use lowest effective dose, monitor QTc
- Dexmedetomidine infusion for agitated delirium in ventilated patients — enables lighter sedation and may shorten delirium duration (limited evidence)
- Avoid physical restraints as a primary strategy — they worsen agitation and delirium
key trials
| trial | question | bottom line |
|---|---|---|
| SPICE III (NEJM 2019) | early dexmedetomidine vs usual care | no mortality benefit; more hypotension/bradycardia. Dexmedetomidine is not superior as a default first-line strategy |
| NONSEDA (NEJM 2020) | no sedation vs light sedation in ventilated patients | no difference in 90-day mortality or ventilator-free days. Non-sedation feasible but not superior |
| MIND-USA (NEJM 2018) | haloperidol vs ziprasidone vs placebo for ICU delirium | no benefit from antipsychotics on delirium duration |
| Kress (NEJM 2000) | daily sedation interruption | ↓ventilator days, ↓ICU LOS |
| ABC (Lancet 2008) | paired SAT + SBT vs SBT alone | ↑survival, ↓ventilator days |
| SLEAP (JAMA 2012) | daily interruption + protocol vs protocol alone | no added benefit — protocol-driven light sedation is sufficient |
what NOT to do
- Benzodiazepine infusions as default sedation — independent risk factor for delirium; reserve for alcohol withdrawal, seizures, benzodiazepine-dependent patients
- Deep sedation without daily reassessment — every day of unnecessary deep sedation adds ~1 day of mechanical ventilation
- Haloperidol for delirium prevention — no evidence of benefit (MIND-USA); QTc prolongation risk
- Restarting sedation at full dose after SAT — restart at half; many patients tolerate much less than they were receiving
- Treating hypoactive delirium with antipsychotics — address underlying causes (infection, metabolic, medications) and mobilise
- Forgetting non-sedative causes of agitation — pain, full bladder, constipation, dyspnoea, ETT malposition