From Evidence to Practice: Managing Pain, Agitation and Delirium in the ICU


Dr. Kim Medlej

Kamal (Kim) Medlej, MD, is an attending physician in the department of emergency medicine at Massachusetts General Hospital and fellowship-trained in critical care, and a longtime contributor to MDCalc. Here’s Dr. Medlej’s take on applying evidence on managing pain, agitation, and delirium to his patients in the ICU.

Pain, agitation, and delirium (the ICU triad) are common in critically ill patients, and can be challenging for clinicians to manage, both in (1) ruling out and treating potential underlying causes, and (2) choosing appropriate sedatives and analgesics in those patients who need them.

This is a fascinating area of critical care medicine, although it can also be difficult and frustrating. While our insight and screening tools have improved, the management of patients with pain, agitation, and/or delirium remains difficult. This is a large topic with a significant amount of literature and opinions. This brief overview is in no way comprehensive, but more a personal view, approach and practice.

In your practice, what’s your general approach to ICU patients with pain or altered mental status? How much of your assessment is based on gestalt, versus using pain scales, delirium scales, and other evidence- or consensus-based tools?

Most ICUs I have worked in have adopted validated screening tools for delirium, such as the CAM-ICU Score. Our nurses usually calculate and document it when assuming care of the patient at the beginning of their shift, and inform the covering physician with any change in condition. It is important to remember that the CAM-ICU Score has limitations, such as in patients that are sedated, or in patients who have baseline confusion or poor cognition. This is  not an insignificant number of patients in the ICU. In these patients, physician and nursing gestalt still play a significant role as does, one must admit, some trial and error.

How do you choose pharmacologic agents for analgesia/sedation? How do you minimize opioid use? Any thoughts on using ketamine in the ICU setting?

This is obviously a complicated question since the patient context is very important. Is the patient intubated? Will the patient need long or short term analgesia/sedation? Is the agitation caused by anxiety/confusion/delirium rather than pain? All of this comes into play when choosing one or a combination of agents. Analgesia and sedation in the early phase of illness is also different than it will be in the days and weeks following the initial insult. Let’s walk through a scenario.

A 63 year old patient presents to the emergency department in acute respiratory distress secondary to community-acquired pneumonia. He is cognitively intact at baseline. He requires intubation and mechanical ventilation. Clinical guidelines recommend an analgesia-first approach to sedation in these patients, as data from a study has found that it increased ventilator free time1. This is also an approach that makes sense since the presence of an endotracheal tube in the patient’s throat rubbing against the oral and pharyngeal mucosa causes a significant amount of discomfort.

In this patient, I would start with an infusion of fentanyl (personal preference given the shorter half life) preceded by a bolus; alternatively, an infusion of hydromorphone preceded by a bolus can also be used. The same data supporting the analgesia-first approach also demonstrated that up to 70% of patients will require an second sedative agent. I have found this to be true in the acute setting, and usually will use two agents when sedating newly-intubated patients.

Studies have shown that the incidence of delirium is lower with propofol and dexmedetomidine compared to benzodiazepines. I tend to prefer propofol and have had good success when combining it with fentanyl. Propofol of course reliably causes hypotension and this can limit its use in some of our sicker patients. I tend to avoid midazolam infusions as much as possible, given the higher incidence of delirium. This is a last recourse for me if I am not able to achieve the level of sedation I need (think intubated asthmatic who is fighting the ventilator) and either the combination of fentanyl/propofol is not enough, I cannot use propofol (patient is too hypotensive), or if there is an indication for the midazolam infusion (status epilepticus).

Dexmedetomidine is a very interesting agent that we are likely to see more of but are currently limited by cost and institutional limitations. A combination of opioids and dexmedetomidine would be ideal as dexmedetomidine does not suppress respiratory drive, and would help identify patients ready to be placed on spontaneous breathing trials and be extubated earlier. The incidence of delirium is also the lowest with dexmedetomidine.

Remifentanil is another agent that we will hopefully see more of in the future. it has a very short half life and does not accumulate with prolonged infusion (unlike fentanyl and hydromorphone).

Finally, ketamine is increasingly being mentioned as an alternative agent for continuous analgesia/sedation in patients that are either intubated or awake and have failed first line therapy. The evidence regarding its use remains poor, with one study reporting a high incidence of adverse events (usually tachydysrhythmias or agitation)2. There is likely a role for in certain subgroups of patients but these have not yet been identified and I think more data is needed before it becomes widely adopted in the ICU. The use of ketamine in sub-dissociative doses for analgesia as a single agent or in combination with an opioid has shown promising results in the emergency department population with a low incidence of side effects. There is less evidence for its use in the ICU, and more studies are needed to demonstrate its safety profile in this patient population at higher risk for agitation and delirium.

So, going back to our patient, I would use a combination of propofol and fentanyl, using higher doses of fentanyl in order to use lower doses of propofol. I would then work on titrating the propofol off and the fentanyl down. There is a strong push to decrease the use of opioids; however, this should not be the priority in the first 48–72 hours, where the focus should instead be on managing the acute illness and/or pain. Once the patient is over the hump, we can start de-escalating the opioid by moving to PRN boluses, a PCA, or even oral medications. The pain assessment should be done on a daily basis and non-opioids such as acetaminophen, NSAIDS, gabapentin, etc, should be introduced as early as possible.

I am personally a big fan of intravenous acetaminophen and have had significant success using it to treat acute severe pain. It remains expensive, however, with many hospitals limiting its use. I certainly hope this will change in the future, especially with the current opioid epidemic.

How do you choose pharmacologic agents for agitation? What are your thoughts on newer agents like Precedex or Geodon over Haldol?

Let us now assume that our patient has been weaned off propofol but remains on a low dose of fentanyl. He is increasingly agitated, to the point that additional sedation is needed to keep him safe from self-extubation. A bedside assessment using the CAM-ICU screening tool is positive for delirium. It is often tempting to dial up the sedation in this already intubated patient, and this would likely help calm him down and stop the monitors and ventilator from alarming. This should only be a last resort, however, as the goal in the ICU is always to move forward with as few setbacks as possible, and by this I mean less sedation, earlier extubation, earlier mobility, etc.

In this situation, I would move to using haloperidol IV in increasing doses, while keeping an eye on the QTc, or olanzapine IV in the acute phase. The goal here is a RASS of 0, rather than true sedation. Alternative agents such as ziprasidone can also be used but have not been found to be more effective or to have a better safety profile. Preference over its use is entirely personal or institutional and I myself have very little experience with it.

Benzodiazepines are one of the main culprits in the development of delirium, and should be only used as a last resort in patients in whom anti-psychotics have failed, or those who are so acutely agitated that they represent a safety risk to themselves or to the ICU staff. In these scenarios, dexmedetomidine has been shown to help reduce the need for intubation and should strongly be considered over benzodiazepines.


  1. Barr J, Fraser GL, Puntillo K, et al. Clinical practice guidelines for the management of pain, agitation, and delirium in adult patients in the intensive care unit. Crit Care Med. 2013 Jan;41(1):263-306.
  2. Umunna BP, Tekwani K, Barounis D, et al. Ketamine for continuous sedation of mechanically ventilated patients. J Emerg Trauma Shock. 2015 Jan-Mar;8(1):11-5.