Sedation, Pain and Delirium Scales

In depth

The one-page summary above is what the PDF prints. These sections go further.

Why analgesia-first light sedation

Pain, anxiety and agitation drive a catecholamine surge: tachycardia, hypertension, higher oxygen consumption, hyperglycemia and catabolism. Deep sedation blunts it at the cost of ventilator days, delirium, pneumonia and ICU-acquired weakness. In an Australian and New Zealand cohort (Shehabi 2012), deep sedation in the first 48 h independently predicted later extubation and higher hospital and 180-day mortality, so day-1 depth matters.

Most 'agitation' in a ventilated patient has a cause: pain, dyssynchrony, hypoxemia, hypercapnia, a full bladder, withdrawal or delirium. Analgosedation means treating pain first (PADIS 2018 good practice statement), usually an opioid titrated to CPOT or BPS, and adding propofol or dexmedetomidine only if the RASS goal is still not met.

Pharmacology and ICU pitfalls

Propofol (GABA-A) works in about a minute and wears off quickly after short infusions, which suits neuro checks and SATs. Vasodilation and myocardial depression make boluses dangerous in hypovolemia. It is a 10% lipid emulsion (1.1 kcal/mL), so count it as nutrition and check triglycerides, commonly at 48 h. Propofol infusion syndrome (metabolic and lactic acidosis, rhabdomyolysis, hyperkalemia, Brugada-like ECG, arrhythmia, AKI) is linked to more than about 5 mg/kg/h for over 48 h, especially with catecholamines and steroids. The label requires changing tubing and opened product every 12 h.

Dexmedetomidine is an α₂-agonist acting at the locus coeruleus. Patients stay rousable and cooperative with little respiratory depression, so it can run through extubation. A loading dose can cause brief hypertension from peripheral vasoconstriction, then hypotension and bradycardia; most ICUs skip it. It does not reliably give amnesia or deep sedation, and stopping after several days can cause rebound tachycardia, hypertension and agitation.

Benzodiazepines remain first-line for alcohol or benzodiazepine withdrawal and seizures but are a modifiable delirium risk factor (PADIS 2018). Midazolam is a CYP3A4 substrate (azoles and macrolides prolong it) with an active metabolite that accumulates in kidney failure. High-dose lorazepam infusions deliver enough propylene glycol to cause an anion gap acidosis with a rising osmolar gap.

Ketamine (NMDA antagonist) gives analgesia while preserving respiratory drive and airway reflexes, and it bronchodilates. It usually raises heart rate and blood pressure, but in a catecholamine-depleted patient its direct myocardial depression can lower pressure. The old teaching to avoid it in brain injury is not supported by current data.

Opioids: fentanyl suits instability and renal failure but accumulates in fat after days of infusion, and large rapid boluses can cause chest wall rigidity. Morphine releases histamine and its active metabolite accumulates in AKI. Hydromorphone is about 7 times as potent as IV morphine, a classic mix-up. After 5–7 days of infusion, taper rather than stop: iatrogenic withdrawal (tachycardia, hypertension, sweating, diarrhea, yawning, agitation) mimics delirium. Many protocols reduce 10–20% a day or convert to enteral methadone; practice varies.

Scale technique, step by step

1. RASS: observe without talking. Calm and alert is 0; restless +1, purposeless movement or dyssynchrony +2, pulling at tubes +3, combative +4.
2. If not alert, say the name and ask the patient to open their eyes and look at you: eye contact over 10 s −1, under 10 s −2, movement without eye contact −3.
3. No response to voice: shake the shoulder, then rub the sternum. Any movement −4, none −5.

Example: eyes closed, opens them to his name, looks at you for 5 seconds and drifts off: RASS −2. Had he turned toward you without eye contact, −3. The first stimulus that gets a response defines the score, so never start with touch.

CPOT: observe 1 minute at rest, then during turning or suctioning, then after any analgesic dose. Four items 0–2. Face: relaxed 0, tense 1, grimacing with eyelids tightly closed 2. Body: no movement 0, protection (cautious movement, touching the painful site) 1, restlessness (pulling the tube, trying to sit up, not following commands) 2. Muscle tension on passive arm flexion: relaxed 0, tense 1, very rigid 2. Ventilator: tolerating 0, coughing but tolerating 1, fighting 2 (vocalization if extubated).

Example: during a turn she grimaces with eyes squeezed shut (2), guards her abdomen (1), resists arm movement (1) and coughs but tolerates the vent (1). CPOT 5, above 2: give an opioid, rescore in 15–30 minutes and premedicate before the next turn.

BPS: three items 1–4. Face: relaxed 1, partly tightened 2, fully tightened with eyelid closing 3, grimacing 4. Upper limbs: none 1, partly bent 2, fully bent with finger flexion 3, permanently retracted 4. Ventilator: tolerating 1, coughing 2, fighting 3, unable to control ventilation 4. Fully tightened face (3), partly bent arms (2) and coughing (2) gives 7, above 5: treat.

CAM-ICU: RASS first; −4 or −5 is 'unable to assess'. Feature 1 comes from the chart and nurse: acute change or fluctuation in 24 h, including a fluctuating RASS. Feature 2: 'Squeeze my hand when I say A', then read SAVEAHAART; a missed A or a squeeze on another letter is an error, and more than 2 is positive. Feature 3: any RASS other than 0. Feature 4: four yes/no questions (Will a stone float? Are there fish in the sea? Does one pound weigh more than two? Can a hammer pound a nail?) and 'hold up this many fingers' (show 2), then 'now the other hand' without showing; more than 1 error is positive.

Example: post-op day 2, RASS 0, pulling at lines overnight (Feature 1 positive), misses 4 A's (Feature 2 positive), RASS 0 (Feature 3 negative), says a stone floats and fails the finger command (Feature 4 positive). CAM-ICU positive.

ICDSC is scored from the whole shift. Consciousness scores 1 for drowsiness or response only to mild-to-moderate stimulation and also for an exaggerated response; a patient responding only to intense stimulation or not at all is not scored. Seven other items score 1 each. Example: drowsy, inattentive and slept under 4 hours scores 3 (subsyndromal); one more item makes 4, delirium.

Delirium: types, risk, prevention and treatment

Inattention is the core of delirium. In 614 medical ICU patients (Peterson 2006), mixed delirium was most common (about 54%), hypoactive next (about 44%) and pure hyperactive rare (under 2%), so most cases look calm or sleepy. Each day of delirium carried about 10% higher risk of death in older ICU patients (Pisani 2009), and longer delirium predicted worse cognition at 3 and 12 months (BRAIN-ICU, Pandharipande 2013).

Prevention is non-drug and bundled: reorientation, glasses and hearing aids, early mobility, night sleep routine, family visits, fewer benzodiazepines, restraints and tethers. PADIS 2018 advises against haloperidol, atypicals, dexmedetomidine, statins or ketamine to prevent delirium. The 2025 update suggests melatonin, which lowered delirium (RR about 0.70).

Treatment starts with the cause: sepsis, hypoxemia, drugs, withdrawal, retention, constipation, pain, sleep loss. MIND-USA (Girard 2018, n = 1,183) found haloperidol or ziprasidone did not change days alive without delirium or coma, ventilator days or survival versus placebo. AID-ICU (Andersen-Ranberg 2022, n = 1,000; haloperidol 2.5 mg IV three times daily, up to 20 mg/day) found no difference in days alive out of hospital at 90 days; lower mortality with haloperidol (36.3% vs 43.3%) was a secondary outcome, not proof of benefit. PADIS 2018 allows short-term antipsychotics for distressing or dangerous symptoms and suggests dexmedetomidine when agitation blocks extubation; the 2025 update made no antipsychotic recommendation.

ABCDEF bundle and SAT/SBT evidence

Each element enables the next: controlled pain (A) allows lighter sedation, paired SAT and SBT (B) shortens ventilation, a non-benzodiazepine choice (C) reduces delirium (D), and an awake patient can mobilize (E) with family involved (F). In 15,226 adults in 68 ICUs (Pun 2019), complete bundle performance was associated with lower hospital death within 7 days (adjusted HR 0.32) and lower next-day ventilation (OR 0.28), coma (0.35), delirium (0.60), restraint (0.37), ICU readmission (0.54) and discharge to a facility (0.64), with a dose-response. These are observational associations.

Daily interruption (Kress 2000, n = 128) cut ventilation from 7.3 to 4.9 days. The ABC trial (Girard 2008, n = 336) paired SAT with SBT against SBT with usual sedation: ventilator-free days 14.7 vs 11.6 of 28, ICU stay 9.1 vs 12.9 days, and 1-year mortality 44% vs 58% (HR 0.68; about 1 life saved per 7 treated). Self-extubation rose, reintubation did not. SLEAP (Mehta 2012, n = 430) added daily interruption to a nurse-driven light-sedation protocol and found no faster extubation but higher midazolam and fentanyl doses: the target is light sedation, and SAT is one way to reach it.

Family engagement means open visiting, joining rounds and care, and early goals-of-care talks; it supports delirium prevention and reduces family anxiety, depression and PTSD (PICS-F).

Sleep promotion and early mobility

ICU sleep is fragmented, spread across the day and short on slow-wave and REM sleep. Propofol and benzodiazepines sedate without normal sleep architecture, and PADIS 2018 advises against propofol to improve sleep. Dexmedetomidine produces an EEG closer to N2 sleep, but PADIS made no recommendation for it as a sleep aid.

• Cluster care for 2–4 hour uninterrupted blocks; move baths, routine labs and weights out of the night.
• Dim lights and noise at night, bright light by day; offer earplugs and eye masks.
• Assist-control rather than pressure support overnight (PADIS 2018) reduces central apneas and arousals.
• Melatonin per the 2025 update; avoid new hypnotics and diphenhydramine.

Schweickert 2009 (n = 104) paired PT and OT with daily sedation interruption early in ventilation: independent function at discharge 59% vs 35%, delirium 2 vs 4 days, ventilator-free days 23.5 vs 21.1. TEAM (Hodgson 2022, n = 750) pushed more intense early mobilization and found no difference in days alive out of hospital at 180 days but more adverse events (9.3% vs 4.1%). The 2025 update still suggests enhanced mobilization: start early and match intensity to the patient.

Alcohol withdrawal in the ICU

Benzodiazepines are first-line, cross-tolerant with alcohol at GABA-A. CIWA-Ar symptom-triggered dosing needs a patient who can talk, so in intubated or delirious patients titrate to RASS. Benzodiazepine resistance is commonly defined as needing more than about 50 mg diazepam in the first hour or 200 mg in 3–4 hours; definitions vary.

Phenobarbital acts at a separate GABA-A site, also blocks glutamate (AMPA), and with a half-life of about 53–118 hours (mean 79, label) tapers itself. It is given as a weight-based load (often 10–15 mg/kg IV, ideal body weight) or as an add-on. In an ED trial (Rosenson 2013, n = 102), one 10 mg/kg dose added to benzodiazepines cut ICU admission from 25% to 8%. Watch additive respiratory depression and hypotension after large benzodiazepine doses; evidence is mostly observational and protocols differ.

Adjuncts: propofol for intubated refractory cases; dexmedetomidine or clonidine for autonomic signs only. Give thiamine before glucose, 200–500 mg IV three times daily if Wernicke encephalopathy is suspected, and replace magnesium, potassium and phosphate.

Post-intensive care syndrome and ICU diaries

PICS is new or worse impairment after critical illness: physical (ICU-acquired weakness, MRC sum score below 48 of 60), cognitive and mental health (anxiety, depression, PTSD); families get PICS-F. In BRAIN-ICU (n = 821), 34% of survivors at 12 months had cognition similar to moderate traumatic brain injury and 24% similar to mild Alzheimer disease.

The bundle is the main prevention. ICU diaries, written by staff and family with photographs, fill memory gaps and replace delusional memories. Jones 2010 (n = 352) found new PTSD at 3 months in 5% with diaries vs 13% without; the larger Garrouste-Orgeas 2019 trial (n = 657) found no reduction in PTSD symptoms. Diaries are low-risk and valued, but the PTSD benefit is unproven. Screen survivors and refer to follow-up clinics where available.

Sedative trials and open questions

In SPICE III (Shehabi 2019) most dexmedetomidine patients still needed propofol or midazolam to reach RASS −2 to +1, and bradycardia and hypotension were more common; a post hoc signal of benefit over age 65 and harm in younger patients is hypothesis-generating only. MENDS2 (Hughes 2021) found days alive without delirium or coma of 10.7 vs 10.8. NONSEDA and Strøm show that no sedation is feasible for some patients but not a default.

Unsettled: the 2025 PADIS update conditionally prefers dexmedetomidine over propofol when light sedation or less delirium is the priority, while the UK A2B trial (Walsh 2025) found α₂-agonists did not shorten time to extubation versus propofol and caused more bradycardia. Optimal depth in early severe ARDS, volatile anesthetic sedation, treatment of hypoactive delirium and the right mobility dose remain open.

Exam pearls

• Analgesia first: a grimacing, vent-fighting patient gets an opioid and a rescore before any sedative.
• CPOT above 2 or BPS above 5 is significant pain; vital signs only prompt an assessment.
• Light sedation is RASS −2 to +1 in the PADIS trials; at −4 or −5, CAM-ICU cannot be done.
• CAM-ICU positive: Features 1 and 2 plus Feature 3 (RASS not 0) or Feature 4.
• Hypoactive and mixed delirium are most common; quiet patients get missed.
• Benzodiazepines are a modifiable delirium risk factor; keep them for withdrawal and seizures.
• Haloperidol does not shorten ICU delirium (MIND-USA, AID-ICU); monitor QTc if used.
• SAT before SBT; failed SAT restarts at half dose; paired SAT and SBT lowered 1-year mortality (ABC).
• Propofol: hypotension, triglycerides, 1.1 kcal/mL, PRIS above about 5 mg/kg/h for over 48 h.
• Dexmedetomidine: rousable, little respiratory depression; bradycardia; not alone for alcohol withdrawal.