Analgesia and Sedation

Analgesia in the ICU

The best assessment of pain control is patient response. However, in critically ill patients who cannot effectively communicate, use of body language or heart rate is probably an ineffective way to measure pain in the critically ill patient. [Crit Care Clin 15: 35, 1999]


Fentanyl is probably superior to morphine because it acts more quickly, has no active metabolites (morphone-3-glucuronide can cause seizures, thus morphine must be reduced in patients with renal insufficiency), and does not produce histamine. Fentanyl’s efficacy is due to its lipid solubility, so if used > 4 hrs fentanyl must be used in the lowest tolerated dose to prevent prolonged effects. PCAs produce better analgesia, better patient satisfaction, and fewer side effects [JAMA 259: 243, 1988] although background dosing is associated with an increased risk for respiratory failure (see below for dosing).

PCA Dosing

  • Fentanyl 15-75 ug q3-10m
  • Dilaudid 0.1-0.5 mg q5-15m
  • Morphine 0.5-3 mg q10-20m

While the data are conflicting, epidurals are generally thought to allow improved analgesia, faster bowel recovery, fewer pulmonary complications, and a decreased risk of myocardial infarction [Anesthesiology 82: 1474, 1995; BMJ 321: 1493, 2000] as compared to systemic opiates but they do not reduce mortality [Anesthesiology 82: 1474, 1995]. Fentanyl is superior to morphine as the incidence of respiratory depression with epidural morphine (1%) is the same as in systemic dosing [Anesthesiology 82: 1474, 1995]. Also, urinary retention occurs in at least 15% (some say as high as 90%).

Other side effects of opiates include respiratory depression, decreases in blood pressure (well tolerated in supine patients [Drug Safety 7: 200, 1992], more of an issue in hypovolemic or heart failure patients or when used with benzodiazepines [Tomicheck RC et al. Anesth Analg 62: 881, 1983])
Opiates cause constipation which may be relieved by naloxone with no systemic effects [J Pain Symp Manag 24: 71, 2002] – “for managing constipation due to opioid use, authors recommend dosages starting as low as 0.8 mg twice daily, with a maximum of 5 mg a day, titrated up to 12 mg a day, watching for toxicity and loss of analgesic effect.13,88,90,91 Particular caution is indicated for physically dependent patients.84 Doses less than 10% of the daily morphine dosage may be ineffective, whereas most patients respond to a naloxone equivalent of 20% of the total daily morphine dosage,88.” In a placebo-controlled, dose-ranging study reported as a letter to the Lancet, the anticonstipation effect of oral naloxone was evaluated in constipated hospice patients receiving morphine or diamorphine by mouth. Oral naloxone, given at a dose of at least 10% that of concurrent opioid analgesia, reversed constipation without reversing analgesia [Lancet 337: 1475, 1991]. Note that giving lower doses more frequently (q4h) may allow adequate first pass degradation while providing the same effect on the GI tract
Pruritis occurs in at least 30% and is often refractory to antihistamines – giving 0.25-1.0 ug/kg/hr of naloxone may abolish this without affecting pain control [J Pain Symp Manag 24: 71, 2002]. Opiates can also cause nausea and vomiting, which are usually relieved by Zofran.

Do NOT use meperidine as nor-meperidine can cause agitation, tremors, myoclonus, delirium, hallucinations, and tonic-clinic seizures [Am J Ther 9: 53, 2002]. The only real use (still unadvised) is to stop post-operative shivering, which it usually does when given 25 mg IV one time dose.


The only IV NSAID available. On a weight-by-weight basis, it is 350 times as potent as aspirin [Drugs 39: 86, 1990]. It takes ~ 1 hour to work and lasts 5-6 hours. Postoperatively, 30 mg is equivalent to 10-12 mg morphine [Drugs 53: 139, 1997], also it has an opioid sparing effect, allowing the dose to be reduced by 25-50% [Drugs 53: 139, 1997]. Note that it is partially metabolized in the liver but also excreted in the urine (reduce to 15 mg q6h in renal failure, age > 65, or weight < 50 kg). Do not give more than 5 days as this increases the risk of GI and operative site bleeding [JAMA 275: 376, 2000]. Ketorolac is known to inhibit platelet aggregation and should never be used in patients at high risk of bleeding [Drug Safety 22: 487, 2000]. Continuous IV infusion of 5 mg/hr is more effective than intermittent IV dosing [Anesthesiology 80: 1277, 1994]

Sedation in the ICU

Single-center study of 128 adults on ventilation with continuous infusions of sedatives in the MICU showed improvements in duration of ventilation (7.3  4.9 days, p = 0.004), ICU stay (9.9  6.4 days, p = 0.02), and mental status when patients were woken up once every 24 hours (beginning 48 hours after initiation of sedation) as opposed to those who were only woken at the discretion of their physician [NEJM 342: 1471, 2000]

Richmond Agitation and Sedation Scale (RASS)[edit]

4 Combative, violent, danger to staff
3 Pulls or removes tube(s) or catheters; aggressive
2 Frequent nonpurposeful movement, fights ventilator
1 Anxious, apprehensive, but not aggressive
0 Alert and calm
-1 Awakens to voice (eye opening/contact) >10 sec
-2 Briefly awakens to voice (eye opening/contact) <10 sec
-3 Movement or eye opening. No eye contact.
-4 No response to voice, but movement or eye opening to physical stimulation.
-5 Unarousable (no response to voice or physical stimulation)


All cause some amnesia. Midazolam, lorazepam, and diazepam can be given IV. In healthy subjects these do not cause respiratory depression, although it can occur in certain ICU patients (ex. those in respiratory insufficiency [Anesthesiology 62: 310, 1985]). Note that while the elimination half-life of midazolam is 2-8 hrs (as opposed to 20-50 for diazepam), the clinical recovery is the same due to lipid solubility of midazolam (also seen in lorazepam) [Crit Care Med 22: 1492, 1994]

Midazolam is the drug of choice for short term use, however one should not use continuous infusion for more than a few hours because it can produce prolonged sedation secondary to fat absorption, hydroxymidazolam metabolites, renal failure, hepatic failure, and P450 inhibition by other drugs. To reduce oversedation risks, dose based on ideal body weight.

Lorazepam has the slowest onset of action of the benzodiazepines and should only be used in patients who require prolonged sedation.

Diazepam is the least favored agent because of the risk for oversedation with repeat administration. It should not be given as a continuous infusion [Crit Care Clin 17: 843, 2001]

Note also that lorazepam and diazepam (but not midazolam) are dissolved in propylene glycol – a bolus may cause hypotension and bradycardia, and prolonged administration can cause a metabolic acidosis, agitation, or a clinical syndrome mimicking sepsis.

Benzodiazepine withdrawal can mimic delirium tremens and is difficult to predict. To reduce the risk of withdrawal in intubated patients on midazolam infusions, consider switching to propofol drip 24 hours prior to extubation as this has been shown to decrease the incidence of post-extubation agitation Saito M ED. Anesth Analg 96: 834, 2003; FREE Full-text at Anesthesia & Analgesia

Lastly, note that diazepam and midazolam are metabolized via hepatic oxidation, whereas lorazepam is metabolized by glucuronidation, thus these two sets of benzodiazepines are affected by different drug classes.


Use in the ICU is limited by the risk of hypotension and decreased myocardial contractility. This is NOT an analgesic. A single bolus works within 1 minute but only lasts 5-8 minutes. It is usually given continuously, after which awakening occurs in 10-15 minutes even after prolonged sedation. Importantly, it has been shown to reduce cerebral oxygen consumption and lower ICP [CNS Drugs 17: 235, 2003]. It has also been used to treat refractory status epilepticus [Ann Pharmacother 32: 1053, 1998] and delirium tremens [Crit Care Med 28: 1781, 2000]. Note that propofol contains 1.1 kcal/mL. It should be dosed based on ideal body weight. Adverse effects include respiratory depression, apnea, and hypotension (especially in patients in shock [CNS Drugs 17: 235, 2003]). Green urine is a rare side effect. Hypertriglyceridemia occurs in 10% of subjects after 3 days of continuous infusion, thus check serum TG in these patients and stop if > 300. In patients given 5 or more days of continuous infusion with AstraZeneca’s preparation, supplement their diet with zinc because AstraZeneca’s formulation contains EDTA. Bradycardia-Acidosis syndrome (triad of bradycardia, hyperlipidemia, and rhabdomyolysis) is rare but lethal in 80% of cases [Paediatr Anesth 8: 491, 1998] – stop the drug if this occurs. Keeping infusion < 4 mg/kg/hr may help prevent this side effect. An interactive, graphical, 3-compartment model of propofol pharmacokinetics is available free of charge at; users can select between the Gepts or Fechner propofol model. The third compartment is the slow compartment that becomes relevant upon long term infusion.


Highly selective alpha-2-agonist that produces sedation, anxiolysis, analgesia, and sympatholysis [Drugs 59: 263, 2000]. Onset within ~ 3 minutes, lasts < 10 minutes. Does not lead to respiratory depression. Usually loaded at 1 µg/kg over 20 minutes (to reduce transient hypertension, found in 15%) followed by 0.2 – 0.7 g/kg/hr continuously. Adverse effects include hypotension (30%), bradycardia (8%), hypertension (15%), and sympathetic rebound if used > 24 hours (thus do not use for more than 24 hours).


Also does not lead to cardiorespiratory depression. Effective in calming patients with delirium. IV route is not FDA approved but supported by the Society of Critical Care Medicine guidelines [Crit Care Med 30: 119, 2002]. D-blocker, works within 10-20 minutes and lasts for several hours. Only causes hypotension in hypovolemic patients or those on beta-blockers. May be ideal for weaning mechanically ventilated patients [Crit Care Med 22: 433, 1994]. IV doses range from 0.5–2 to 5-10 to 10–20 mg for mild, moderate, and severe anxiety. Individual patients show significant variation in serum levels after a single dose, so if there is no response give a second, double-dose after 10 minutes. Do not give a third dose (switch agents). Extrapyramidal reactions are rare and are decreased in incidence when benzodiazepines are added [J Intensive Care Med 4: 201, 1989]. More feared adverse events include NMS and torsades – note that 3.5% of patients on Haldol will show QT prolongation [Am J Cardiol 81: 238, 1998], thus pre-existing QT interval is a contraindication. Haldol-associated QT prolongation may be exacerbated in the presence of class III antiarrhythmics, hypocalcemia and intracranial hypertension. Also, Haldol is a mild selective -antagonist. Its safety has been questioned in acute head injury, as animal studies suggest worsening of secondary brain injury by the central antidopaminergic effect – a recent animal study suggested that single or multiple low doses of risperidone and haloperidol may be innocuous to recovery after traumatic brain injury, but that chronic high-dose treatments are detrimental [Crit Care Med 35:919, 2007]. Lastly, Haldol may lower the seizure threshold (Andrews, citation needed).


Beware anticholinergic effects. Often used for sedation in pediatric dental patients, may not be more effective than placebo.


Sometimes used for alcohol withdrawal as well as to decrease pain transmission.