Staying woke IV: Delirium of critical illness

Previously in our discussion of ICU liberation, we discussed The Basics, Sedation, and Mobility. Now let’s dive further into the topic of delirium.

What is delirium, and why does it afflict many of our patients?

Delirium is a “disturbance of consciousness characterized by acute onset and fluctuating course of inattention accompanied by either a change in cognition or a perceptual disturbance.” In other words, delirious patients demonstrate abnormal levels of attention and arousal and warped, incorrect comprehension and perception. Importantly, this is a new change in mental status resulting from an acute illness, in contrast to chronic conditions such as dementia. People get sick and then their brains stop working correctly.

How does this look? Classically, the delirious patient is agitated and confused, combative, and hallucinating — he may believe himself to be abducted by aliens, or trapped inside a well, or undergoing physical torture. His behavior may be appropriate within that context, which often means restlessness or outright combativeness. Sedation may be needed.

This is hyperactive delirium, a familiar phenomenon in the ICU. However, another phenotype is at least as common: hypoactive delirium. Here, awareness and cognition are just as impaired, but arousal is muted, not elevated. These patients are not thrashing and yelling, but are lying quietly in bed, staring at the wall. If a careful evaluation is undertaken, their thoughts are just as disordered — but superficially, they aren’t obviously ill. In fact, they’re a model patient. In reality, though, the implications of their delirium are just as serious.

Clearly, delirium is more of a syndrome than a singular entity. Essentially, it’s simply an acute dysfunction of the brain, similar to the other organ injuries we often see in the setting of critical illness: acute renal insufficiency, transiently elevated troponins, and so on. Its causes are complex, and any attempt to peek inside the mechanisms reveals a tangled web of pathways we are only beginning to understand.

But its consequences seem clear.

Harms of delirium

In numerous datasets — mostly without randomized controls, because we cannot randomize patients to delirium, but nevertheless a substantial body of observational data — delirium has been independently associated with mortality. In other words, after controlling for every other variable we can, ICU patients who become delirious seem to die more often than otherwise identical non-delirious patients. Depending on the study, this effect may be as large as:

Again, this is observational data. It’s clear that delirium is often a consequence of critical illness, and sicker patients are more likely to be delirious. But despite statistically balancing these confounders to the best of our ability, still the effect seems to remain. Patients who get delirious are more likely to die.

They also stay in the ICU longer. Development of delirium independently prolongs both ICU and hospital stay by as much as 5–12 days (1, 2, 3) depending on the study. In some data it was actually the single strongest predictor of length of stay.

Finally, delirium is independently associated with prolonged mechanical ventilation, stretching out vent time by as much as 10 days (1, 2).

In short, delirium is right at the heart of the problems we’re trying to prevent in the post-resuscitation phase of care, promoting and embodying the dysfunction that leads to more complications and impedes recovery.

Diagnosing delirium

It’s clear that without active screening, delirium is wildly underdiagnosed. If you only catch it when the patient is yelling the diagnosis at you, you’re missing many cases; you need to go look for it. While in principle this diagnosis could lie in domain of a psychiatrist or rehabilitation psychologist, in practice the critical care service is usually the best team (and often the only team) to address it.

Two validated, purpose-built screening tools have been developed for ICU delirium: the Confusion Assessment Method for the ICU (CAM-ICU) and Intensive Care Delirium Screening Checklist (ICDSC). The former seems to have emerged as the most popular. These simple bedside assessments can be readily incorporated into the workflow of any ICU, with nurses performing them on every patient as often as multiple times per day. Positive results can be conveyed to the treating team.

What on earth should we do about it?

Preventing and treating delirium

We can answer this simply: there is no treatment for delirium.

At least, there is no disease-specific therapy that’s proven to quickly and easily exorcise it. The one possible exception is dexmedetomidine (Precedex), which emerging data suggest may accelerate resolution of delirium. Maybe. Otherwise, sedatives are sometimes needed for safety, but whether you use propofol, antipsychotics, or benzodiazepines, you’re merely treating the symptoms, not the disease — and in some cases you may be exacerbating it, since sedation itself is deliriogenic. (Don’t fall into the thinking that neuroleptics, such as haloperidol or quetiapine, are a “treatment” for delirium because of their role in schizophrenia. They’re just another sedative — perhaps a better choice of sedative than a benzo, but they won’t help normalize the abnormalities of thought and perception.)

For the most part, the only treatments are time, healing, and deescalation. The triggers and risk factors that contributed to delirium must be removed and the patient must be allowed to get better. The diagnosis of delirium should prompt a review to ensure that you’re doing everything possible to avoid delirium-promoting practices — something we ought to be doing anyway. But like checking your speedometer when you see a police cruiser, sometimes we need the reminder.

Many of those triggers are simply the things we’ve discussed already. Limiting sedation and choosing the best agents. Mobilizing patients as early as possible. What else can be done?

Here are a few extra measures that might make a difference.


Sleep in the ICU is difficult due to noise, interruptions, and the alien environment. The resulting sleep deprivation may predispose patients to delirium. A normal sleep/wake cycle should be promoted, allowing uninterrupted sleep overnight to the greatest extent possible, and matching wakefulness during the day; once this cycle becomes deranged (with frequent sleep during the day and insomnia at night), delirium is often close at hand. Basic measures to promote the sleep/wake cycle can include:

  • Managing light. Curtains and blinds should be closed at night and opened in the morning. Room and hall lights should be dimmed overnight. Disposable eyemasks can be offered to non-delirious patients.
  • Managing noise. Earplugs can be offered to non-delirious patients, and efforts made to curb staff chatter, overhead announcements, and other night-time sounds.
  • Managing interruptions. Unnecessary lab draws, vital checks, and other care overnight should be rescheduled, consolidated, or eliminated whenever possible.
  • Basic pharmacologic interventions. Sleep is not the same as sedation. However, carefully selected oral hypnotics may have a role for some patients, such as trazodone, zolpidem, or quetiapine.

While the data is not yet well developed, Kamdar 2013 did find that a sleep package of this type increased ICU days without delirium (from 48% to 43%) and reduced overall risk of delirium/coma from 69% to 49%. Similarly, Van Rompaey 2012 found that simply offering earplugs and eyemasks early in an ICU stay reduced confusion/delirium by 53%.

Family involvement

Patients may become delirious less often when they have the familiar faces of family and friends at the bedside — and when they do, they may have better outcomes with these people present to redirect them, reorient them, and monitor their safety. Rosa 2017 found that the simple measure of extending an ICU’s permitted visiting hours reduced delirium from 20.5% to 9.6%, the duration of delirium/coma by 1.5 days, and ICU length-of-stay by 1 day.

This costs nothing and many families are happy to do it. The concept can be extended even further, however, by asking those family members to participate in patient care to the extent of their ability and willingness (Wyskiel 2015). This can include basic tasks like repositioning, oral care, and sensory stimulation (reading, playing music, watching TV), and while not proven to improve outcomes, can certainly lighten the nursing workload.

Wyskiel family involvement “menu”
Sensory stimulation with occupational therapy

Introducing controlled, graded sensory stimulation into their environment has the potential to improve recovery of neurologic function in brain-injured patients. This practice is usually implemented by occupational therapists with the help of nursing and other bedside staff, and while the data on “coma stimulation” is still early, it is encouraging.

Along with performing the usual gross motor training in conjunction with physical therapy, this sort of activity is a good niche for occupational therapists — focusing on cognitive retraining, reintroducing activities of daily living to force patients to utilize skilled behaviors and normalize their experiences, and helping them communicate and reorder their thoughts.

Facilitate communication with speech language pathology

I used to bang on things, you know? Frustrated as hell. . . . Oh, I was really getting mad. . . . I’m thinking to myself, “why can’t they understand me?” . . . it’s a punishment, I think. It’s terrible. [Caroll 2007]

Voicelessness and the inability to communicate is repeatedly cited in qualitative studies as one of the worst aspects of the ICU experience (1, 2) — potentially worse than pain or illness itself. Yet it is widely experienced by our intubated or trached patients. Anything we can do to help them communicate is not only a kindness, it likely reduces the need for sedation and the perception of confusion. How many nurses and providers would look at the patient quoted and decide to knock him out?

Speech pathologists are often the best people to facilitate this. They can fit speaking valves (“Passy Muir” valves) onto tracheostomies and provide communication boards, eye boards, and writing pads for intubated patients. Newer devices have expanded these offerings, such as in-line speaking valves for trached patients still on the ventilator, and even “talking trachs” that permit vocalization despite a cuffed , fully inflated trach.


  1. Delirium in the critically ill is independently associated with morbidity and mortality.
  2. It is underdiagnosed if specific testing is not performed; such screening (using CAM-ICU or similar scores) should be done routinely on all ICU patients.
  3. There are no evidence-based treatments for delirium, with the possible exception of dexmedetomidine, but prevention and perhaps acceleration of its resolution can be achieved by sedation management, early mobility, promoting physiologic sleep, engagement with family, facilitation of communication, and basic cognitive interventions.

Next time we’ll wrap things up with some basic recommendations for how to apply these concepts in your ICU and improve outcomes among your own patients.

One Reply to “Staying woke IV: Delirium of critical illness”

Leave a Reply

Your email address will not be published. Required fields are marked *