Delirium is a disturbance of consciousness and cognition that develops acutely (ie, hours to days), fluctuates over time, and is generally reversible.1 It is characterized by an acute change or fluctuation in baseline mental status, inattention, and either disorganized thinking or an altered level of consciousness. ICU delirium is the most common form of acute brain dysfunction in critically ill patients, with estimates of incidences ranging from 20% to 50% in nonventilated patients and 60% to 80% in ventilated ICU patients, depending on the diagnostic criteria used and the patient characteristics.2,3,4,5,6 It is important to recognize that delirium is not a normal part of critical illness, but rather represents an acute organ failure that is associated with profound short- and long-term consequences.
While the pathophysiology of delirium is still poorly understood, it is thought to be a disease-driven process caused by the complex interaction of various factors including (1) the underlying disease itself, (2) predisposing risk factors unique to each patient, and (3) environmental and treatment-related factors.
Delirium is hypothesized to be caused by imbalances in neurotransmitters due to factors such as systemic inflammation, metabolic derangements, acute stress responses, and exposure to psychoactive medications.7 The 2 main neurotransmitters implicated in these derangements are dopamine and acetylcholine, and they work in opposition by increasing and decreasing neuronal excitability, respectively. Other neurotransmitters that have been implicated in the pathogenesis of delirium include g-aminobutyric acid (GABA), serotonin, glutamate, and endorphins. At this time, data supporting these hypotheses are limited and thus pharmacologic treatment of ICU delirium is largely empirical.
Inflammatory cytokines and endotoxins are postulated to contribute to the development of ICU delirium in several ways.8 Infiltration of leukocytes and cytokines into the central nervous system may lead to neuronal apoptosis, and may interfere with neurotransmitter synthesis and neurotransmission, increase vascular permeability, and reduce cerebral microvascular blood flow through the formation of microaggregates of fibrin, platelets, erythrocytes, and neutrophils. Microglial activation and oxidative injury can also occur.
Alterations of Brain Structure
Preliminary studies using magnetic resonance imaging have shown an association between the duration of ICU delirium and both cerebral white-matter disruption and cerebral atrophy. Because of the lack of imaging before critical illness, these findings suggest that either the presence of these abnormalities makes patients more vulnerable to developing ICU delirium or that ICU delirium leads to abnormalities in brain structure.9,10
Despite our relatively poor understanding of the overall pathophysiology of delirium, clinical studies have shown delirium to be a disease-driven process caused by a complex interaction of factors including: (1) baseline risk factors unique to each patient; (2) the patient's acute illness and illness-related factors; and (3) ICU-environment and treatment-related factors (Figure 49–1). These risk factors can be further classified as nonmodifiable (ie, predisposing factors that are out of the control of the admitting clinician) and modifiable (ie, factors that the clinician may be able to treat). Although over 100 risk factors have been identified in the literature,11 few have consistently remained associated with delirium across different studies after adjusting for confounding variables. This is likely due to the different underlying pathophysiology of delirium across study populations.2,3,6,12,13,14 This chapter will review risk factors that have been most consistently identified in the literature.
Risk factors for ICU delirium.
Copyright © 2002, E. Wesley Ely, MD, MPH and Vanderbilt University, all rights reserved (top).
(Data from Devlin JW, Marquis F, Riker RR, et al: Combined didactic and scenario-based education improves the ability of intensive care unit staff to recognize delirium at the bedside, Crit Care. 2008;12(1):R19 (bottom).)
Patient-Level Predisposing Risk Factors
Studies have consistently identified pre-existing cognitive impairment, alcohol use, and history of hypertension as risk factors that significantly increase the risk of delirium in ICU patients (Figure 49–2).2,5,13,15 Although advanced age has been identified as one of most significant risk factors outside of the ICU, this association has remained inconsistent across ICU studies.16,17 Physiologic dependence from chronic exposure to alcohol, opiates, and benzodiazepines can lead to withdrawal in the setting of abrupt discontinuation of alcohol use. Of the predisposing patient-level risk factors, withdrawal from these substances is the only modifiable factor, and thus should be carefully considered when encountering a delirious patient.
Delirium assessment tools. (Reproduced with permission from Hsieh SJ, Ely EW, Gong MN: Can intensive care unit delirium be prevented and reduced? Lessons learned and future directions, Ann Am Thorac Soc 2013 Dec;10(6):648-656.)
Illness-Related Risk Factors
High severity of illness at ICU admission and medication-induced coma (as opposed to coma due to a primary neurologic condition) have been consistently identified as independent risk factors for development of delirium in ICU patients.5,13,16,17,18,19 Respiratory failure requiring mechanical ventilation has been identified as a risk factor for delirium, although the results have been inconsistent across studies.2,5,14,16,20 Potential reasons for the discrepant findings include exclusion of nonmechanically ventilated patients from observational studies12,19,21 and lack of measurement of mechanical ventilation as a risk factor in ICU delirium studies.13,15
ICU Treatment-Related Risk Factors
An appreciation of the impact of ICU-level risk factors on the development and persistence of delirium is particularly important because (1) these factors are all modifiable, (2) reduction of these risk factors is associated with reduced incidence and duration of ICU delirium and improvements in clinical outcomes, and (3) these factors are closely interrelated.
Sedative Use—Most mechanically ventilated patients receive sedatives in the ICU. Both sedative choice and sedative-induced coma are independently associated with an increased risk of delirium.13 Benzodiazepines are most consistently associated with delirium across different ICU populations and have demonstrated a dose-dependent relationship, although a few studies have found no significant relationship.12,16,17,22,23,24 Most studies on opiates and propofol report an increased risk of delirium, particularly when used in combination with other sedatives or when associated with coma.2,13,14,16,25 In contrast, recent studies suggest that use of dexmedetomidine and/or avoidance of benzodiazepines may be associated with both a lower risk of developing delirium and a shorter duration of delirium.26,27 Preliminary data in cardiac surgery patients suggests that dexmedetomidine may also be associated with a lower incidence of delirium compared to propofol and shorter duration of delirium compared to morphine.28,29
Immobility—A number of observational studies suggest that neuromuscular function and ICU delirium are closely interconnected.30 For instance, observational and clinical trial data suggest that immobility is an independent risk factor for delirium in ICU and non-ICU hospitalized patients.5,22,30,31,32,33 Studies show that ICU's that institute early mobilization programs have a lower prevalence and shorter duration of ICU delirium.33,34
ICU delirium is characterized by the following 4 DSM-V criteria: (1) inattention (the most common feature), (2) an acute change or fluctuation in baseline mental status, and either (3) disorganized thinking or (4) an altered level of consciousness.1 Notably, while delusions and hallucinations can be present in delirious patients, these symptoms are not defining features of delirium. Patients with delirium can present in 3 different ways: calm or somnolent (ie, hypoactive), agitated (ie, hyperactive), or alternating between the 2 states (ie, mixed).24 While hyperactive delirium can be more easily identified without a screening tool, only 2% of patients have purely hyperactive delirium. In contrast, hypoactive and mixed forms are much more common in critically ill patients,35 and are associated with worse clinical outcomes.35
ICU delirium is a clinical diagnosis; no diagnostic lab, imaging, or electroencephalographic test can accurately diagnose delirium. However, without the use of a delirium assessment tool, over 3/4 of delirium can be missed in routine practice because patients more often present with hypoactive rather than hyperactive delirium.36 This highlights the importance of integrating a structured tool into clinical practice to rapidly and accurately detect delirium, rather than relying on clinical impressions. Indeed, delirium assessment in the ICU is now considered to be a requisite part of high-quality ICU care.
All patients should be screened for delirium as soon as they are admitted to the ICU and then at least once per nursing shift. Because delirium can only be assessed in patients who are arousable to voice, level of consciousness needs to be determined first. Therefore, delirium screening is a 2-step process.
Step 1: Determine Level of Consciousness—The level of consciousness needs to be determined using a sedation-agitation scale. The 2013 American College of Critical Care Medicine (ACCM) clinical practice guidelines recommended the Richmond agitation-sedation scale (RASS)37 and Riker sedation-agitation scale (SAS)38 as the 2 most valid and reliable sedation assessment tools (Table 49–1 and 1B).
Table 49–1A ||Download (.pdf) Table 49–1A
|Richmond Agitation-Sedation Scale (RASS)88 |
|+4: Combative ||Overtly combative, violent, immediate danger to self |
|+3: Very agitated ||Pulls or removes tubes or catheters; aggressive |
|+2: Agitated ||Frequent nonpurposeful movement, fights ventilator |
|+1: Restless ||Anxious but movements not aggressive or vigorous |
|0: Alert and calm ||Alert and calm |
|–1: Drowsy ||Not fully alert but has sustained awakening to voice (eye opening or eye contact >10 s) |
|–2: Light sedation ||Briefly awakens with eye contact to voice (< 10 s) |
|–3: Moderate sedation ||Movement or eye opening to voice but no eye contact |
|–4: Deep sedation ||No response to voice but movement or eye opening to physical stimulation |
|–5: Unarousable ||No response to voice or physical stimulation |
Table 49–1B ||Download (.pdf) Table 49–1B
|Riker Sedation-Agitation Scale (SAS)89 |
|7: Dangerous agitation ||Pulling at endotracheal tube, trying to remove catheters, climbing over bed rail, striking at staff, thrashing from side to side |
|6: Very agitated ||Requiring restraint and frequent verbal reminding of limits, biting endotracheal tube |
|5: Agitated ||Anxious or physically agitated, calming at verbal instruction |
|4: Calm and cooperative ||Calm, easily arousable, follows commands |
|3: Sedated ||Difficult to arouse but awakens to verbal stimuli or gentle shaking; follows simple commands but drifts off again |
|2: Very sedated ||Arouses to physical stimuli but does not communicate or follow commands, may move spontaneously |
|1: Cannot be aroused ||Minimal or no response to noxious stimuli, does not communicate or follow commands |
Step 2: Screen for Delirium—If the patient is not comatose (ie, RASS –3 or more or SAS 3 or more), delirium can be assessed using a screening tool. Many tools for delirium screening have been published. This chapter will review the 2 most well-studied and commonly used ICU delirium screening tools, that were also recommended by the recently updated ACCM clinical practice guidelines: the Confusion Assessment Method-ICU (CAM-ICU) and Intensive Care Delirium Screening Checklist (ISDSC) (Figure 49–1).24,40 Both tools have (1) high sensitivity, specificity (ranging from 74% to 96%) and (2) excellent clinical feasibility in both nonventilated and ventilated critically ill patients in mixed ICUs.41
Confusion Assessment Method-ICU (CAM-ICU)
The CAM-ICU provides a dichotomous assessment (ie, delirious vs not delirious) at a single time point and can be performed in less than 1 minute. Advantages of the CAM-ICU are its discrete defined measures that are obtained from physical assessment of the patient. A disadvantage is that, given delirium's fluctuating course, the periodic nature of the CAM-ICU may miss an episode of delirium. Therefore, it should be performed on a regular basis (eg, every 4 to 12 hours) and with changes in the patient's mental status.
The ICDSC is an 8-item checklist of symptoms observed over an 8-hour to 24-hour period in which a score of 4 or more is positive for delirium and scores of 1 to 3 are defined as “subsyndromal” delirium. Advantages of the ISDSC are its ability to detect “subsyndromal delirium” and a longer assessment period which decreases the chance of missing signs of delirium. A disadvantage is that the ISDSC relies on more subjective observations in the setting of mechanically ventilated patients (eg, hallucinations, inappropriate speech) and thus can be more dependent on the clinical experience of the practitioner.
Untreated pain can be both a risk factor for ICU delirium and a cause of agitation that is not due to delirium. While the standard for pain assessment is self-report, patients in the ICU often are unable to communicate due to respiratory failure or decreased level of consciousness. Because increased vital signs such as tachycardia or hypertension do not always correlate with pain, it is important to use a structured tool for pain monitoring in patients who are unable to communicate, such as Behavioral Pain Scale (BPS)42 or the Critical-Care Pain Observation Tool (CPOT).24,43
While dementia is a risk factor for delirium, it can also be confused with delirium because of changes in cognition. However, its course progresses over a much longer period of time (months to years) in contrast to hours to days in delirium, and the symptoms fluctuate much less. In addition, attention remains relatively intact whereas inattention is the most common feature of delirium.
Agitation due to underlying psychosis or mania, and inattention due to underlying depression can masquerade as delirium. However, the symptoms associated with psychosis, mania and depression can persist for longer periods of time and fluctuate less, and the sensorium is usually clear.
Data on effective pharmacologic prevention strategies are limited. A few preliminary studies in patients undergoing elective surgery suggest that low-dose haloperidol and low-dose risperidone reduced the incidence of postoperative delirium.23,44 Because these studies involved patients with a low severity of illness, it is unclear if these findings can be extrapolated to the general ICU population. Dexmedetomidine may be associated with less incident postoperative delirium in cardiac surgery patients, compared to propofol or midazolam.28 However, these findings need to be confirmed before broad treatment recommendations can be made.
Several reasons might explain why studies on successful delirium prevention strategies are limited to date. First, up to 70% patients are admitted to the ICU with delirium already present.16 Second, prevention studies require a larger sample size to detect a differences in incident delirium (which is a binary outcome) compared to duration of delirium (which is a continuous outcome). Finally, most pharmacologic delirium prevention studies did not include concurrent nonpharmacologic delirium prevention strategies such as sedation titration and early mobilization. Because delirium is multifactorial in origin, addressing only a few of the many factors contributing to its development may limit the efficacy of a prevention strategy.
In contrast, preliminary studies suggest that nonpharmacologic strategies targeting multiple risk factors for delirium may be more effective. These strategies include early rehabilitation, sleep-promotion, and structured reorientation. Because these prevention strategies overlap with treatment strategies, they will be discussed in greater detail in the following ICU-level treatment strategies section.
Several high-quality intervention studies to halt and limit the duration of delirium in its earliest stages have successfully reduced the duration of delirium and improved other clinical outcomes. This suggests that despite the occurrence of delirium, strategies to reduce the delirium duration may be the first area of focus for ICU teams. In order to maximize the therapeutic potential, a combination of strategies should be used: (1) multicomponent nonpharmacologic interventions that are useful for all ICU patients should be implemented at an ICU level (ICU-level strategies) and (2) pharmacologic interventions that may be useful for specific patients and should be titrated to each individual (patient-level strategies) (Figure 49–3). Each of these strategies will be discussed in the following sections.
A clinically useful approach to ICU delirium prevention and treatment. This proposed approach incorporates patient-level delirium prevention and reduction assembled from multiple evidence-based sources, but has not yet been tested in a critically ill population. (Reproduced with permission from Hsieh SJ, Ely EW, Gong MN: Can intensive care unit delirium be prevented and reduced? Lessons learned and future directions, Ann Am Thorac Soc 2013 Dec;10(6):648-656.)
Given the multifactorial nature of delirium and the interdependency of ICU-treatment–related risk factors, it is not surprising that multicomponent ICU-level strategies have had better success with reducing the duration of delirium compared to pharmacologic strategies that address only a few ICU-level risk factors (Figure 49–4).
ICU-level delirium prevention and reduction strategies are interconnected. (Reproduced with permission from Hsieh SJ, Ely EW, Gong MN: Can intensive care unit delirium be prevented and reduced? Lessons learned and future directions, Ann Am Thorac Soc 2013 Dec;10(6):648-656.)
Nearly 50% of critically ill patients experience significant pain during their ICU stay.42 Studies suggest that pain may be a risk factor for delirium.12,13 Several possible reasons to explain this relationship are: (1) the deleterious cognitive effects of pain itself, (2) agitation due to untreated pain leading to inappropriate sedative administration, and (3) pain medication doses in excess of what is required for pain control. Indeed a study showed that patients who are regularly assessed for pain received less sedation compared to those who did not receive regular pain assessments.45 Therefore, the goal of pain management through routine pain monitoring should be satisfactory pain control without oversedation, and pre-emptive treatment of pain before painful procedures are initiated.
Agitation is common in critically ill patients. Determining the underlying cause(s) of agitation is essential for determining the appropriate treatment for agitation. Common causes for agitation include untreated pain, anxiety, withdrawal from alcohol or chronic opiates or sedatives, factors associated with the acute illness (eg, hypoxia and hypotension), and delirium. Treatment of agitation that does not address the underlying cause can incite or prolong delirium (eg, a patient with agitation due to uncontrolled pain is given benzodiazepines). Conversely, successful treatment or responsiveness to the underlying cause can potentially reduce sedation use and even improve clinical outcomes.
Medication-induced coma is a risk factor for ICU delirium. In addition, the prolonged immobility that mechanically ventilated patients experience during deep sedation can lead to complications such as muscle atrophy and weakness, ventilator dependency, pressure sores, and venous thromboembolic disease.46,47,48 Rather than routinely providing deep sedation to mechanically ventilated patients without a specific indication, the overall goal of sedation should be to achieve a level of wakefulness so patients can actively participate in rehabilitation and cognitive stimulation during their critical illness. Three different approaches that have demonstrated good outcomes are targeted: (1) daily interruption of sedation, (2) light sedation, and (3) no sedation. Strategies to decrease sedation have led to decreased delirium duration, reduced time on the mechanical ventilator, decreased ICU length of stay, and decreased mortality, and have been demonstrated to be safe, feasible for incorporation into daily care, and acceptable to ICU staff.49,50,51,52,53,54 At least one, if not all, of these approaches should be adopted for a “less is more” culture of sedation use.24,55 Routine monitoring of quality and depth of sedation is needed to guide these strategies. The Richmond agitation-sedation scale (RASS)37 and sedation-agitation scale (SAS)38 were identified by the ACCM Clinical Practice Guidelines as the 2 most valid, reliable, and feasible sedation assessment tools for goal-directed sedation delivery.24
Multiple studies have consistently identified immobility as a risk factor for delirium.5,31 Furthermore, ICU delirium is associated with functional disability after hospital discharge.56,57 Early delivery of physical and occupational therapy to mechanically ventilated ICU patients (eg, passive range of motion in unresponsive patients, active exercises in interactive patients) has been associated with reduced delirium prevalence and duration, and is safe and well-tolerated.33,58 In addition, early rehabilitation is associated with less time on mechanical ventilation, improved return to functional status, and may lead to cost savings.59 Of note, studies demonstrating the benefits of early mobilization also targeted other ICU delirium risk factors such as sedation reduction coordinated with mechanical ventilator weaning.
Critically ill patients frequently experience poor sleep quality in the ICU.60,61,62 Sleep deprivation has been postulated to be a risk factor for ICU delirium because both sleep-deprived and delirious patients share common clinical and physiologic derangements.63 Preliminary studies suggest that a combination of nonpharmacologic sleep promoting interventions (eg, earplugs and reducing nighttime procedures and noise, daytime mobilization), and decreased use of sedatives known to alter sleep or precipitate delirium (ie, benzodiazepines, opiates, diphenhydramine, trazodone), can improve self-reported sleep quality and may even reduce incident delirium and delirium duration.64,65,66,67 Although more work is needed in general ICU populations, given the relative ease of implementation, minimal risk, and potential benefit of these interventions, it would be reasonable to implement these practices into usual care.
Pre-existing cognitive, visual, and hearing impairment are risk factors for delirium, likely because of the disorientation that patients with these impairments experience.5,15 Reorientation strategies, such as reading newspapers, listening to music, wearing visual and hearing aids, and performing cognitively stimulating activities, have effectively prevented delirium in older non-ICU patients; preliminary studies in ICU patients also suggest a benefit.68,69
Patient-Level Pharmacologic Treatment
Data on pharmacologic treatment of ICU delirium is mixed. Current evidence suggests a potential benefit from dexmedetomidine and antipsychotics, but other agents such as cholinesterase inhibitors and melatonin have not been found to be helpful in preventing delirium and may even be harmful in the case of cholinesterase inhibitors.70,71,72 The limited success of these therapies in clinical trials may in part be due to the lack of concurrent nonpharmacologic multicomponent delirium prevention strategies such as early rehabilitation and reduced sedation practices.
Dexmedetomidine is a selective a2-adrenoreceptor agonist that has sedative, analgesic and anxiolytic properties. Studies suggest that it is associated with less delirium and may promote better sleep/wake cycle regulation when compared to medications that work through the GABA receptor pathway such as benzodiazepines.73 Several large, well-designed randomized controlled trials comparing dexmedetomidine vs benzodiazepines for sedation in mechanically ventilated medical and surgical ICU patients have shown that dexmedetomidine was associated with a 30% lower prevalence of delirium26 and more days alive without delirium and coma (7 vs 3 days).27 In addition, patients receiving dexmedetomidine spent less time on mechanical ventilation (3.7 vs 5.6 days),26 and dexmedetomidine was not associated with increased cost.27 While evidence comparing dexmedetomidine to other sedatives such as opiates or propofol in the general medical and surgical ICU patient population are more limited,74 these data are encouraging, particularly since the benefit was observed even in the setting of good sedation practices (eg, daily sedation vacation, targeted light-moderate sedation level, delirium monitoring).
While data are currently insufficient to support the widespread use of dexmedetomidine for sedation in all ICU patients, the 2013 ACCM guidelines recommend that dexmedetomidine could be considered for use as a sedative in patients with delirium and in patients who are at high risk for delirium.24
Antipsychotics (eg, haloperidol, risperidone, quietiapine) are hypothesized to treat delirium by blocking dopamine-mediated neuronal excitability and thus stabilizing cerebral function.75 Although they were formerly recommended by major guidelines for treatment of ICU delirium (and are still widely used for that indication),76,77 no large-scale prospective RCTs have tested the impact of antipsychotics on delirium duration, and the 2013 ACCM guidelines no longer recommend for, or against, their use. A small trial in ICU patients with delirium who were already receiving haloperidol found that delirium resolved faster in patients who received haloperidol plus quetiapine, compared to patients who only received haloperidol.78 Clinical trials are currently underway to determine if antipsychotics are an effective treatment for ICU delirium.
Delirium is a strong predictor of ICU length of stay, even after adjusting for factors such as severity of illness and age,79 and is independently associated with poor short-term consequences including increased duration of mechanical ventilation, increased hospital length of stay, and institutional placement.4,80,81,82,83 While even 1 day of delirium is associated with poor clinical outcomes, it is also important to recognize that a “dose-dependent” relationship exists between the duration of delirium and poor clinical outcomes. For each day a patient is delirious, the risk of 6-months and 1-year mortality increases by 10%.4,82 In addition, a longer duration of delirium is an independent predictor of cognitive impairment in both older and younger mechanically ventilated patients.83,84 A recent study found that up to 34% of patients had persistent deficits in global cognition and executive function that were similar to mild Alzheimer's disease and moderate traumatic brain injury 1 year after critical illness.85 Increased delirium duration is also associated with disability in activities of daily living and worse motor-sensory function in the year following critical illness.86 These adverse consequences can profoundly impact a patient's ability live independently after hospital discharge and can decrease their health-related quality of life. It can also be highly distressing for family members and caregivers and increase their caregiver burden.87 With an annual cost of $4 to $16 billion in the United States alone,81 ICU delirium is now recognized as a major public health problem.
Current Controversies and Unresolved Issues
Over the last 10 years, significant advances have been made in understanding risk factors for ICU delirium and have resulted in effective ICU-level strategies that have reduced the adverse impact of delirium. Nonetheless, many questions remain. First, animal models and trials on pathway modulation are needed to elucidate the pathophysiology of delirium. Second, the optimal pharmacologic therapy to prevent and reduce ICU delirium is still unknown, and the optimal protocols for different patient populations still need to be determined. Third, it is unclear if the improved short-term clinical outcomes that are associated with delirium reduction (eg, decreased time on mechanical ventilation, decreased ICU length of stay) translate into improved long-term cognitive, functional, and psychological outcomes. Fourth, more work is needed to elucidate the clinical implications of delirium severity and its subtypes (eg, subsyndromal delirium; hypoactive vs hyperactive delirium). Finally, more studies on cognitive and physical rehabilitation are needed to determine the optimal prescription, timing, and duration for different ICU patient populations.