|Home | About | Journals | Submit | Contact Us | Français|
Delirium is an acute change in cognition and attention, which may include alterations in consciousness and disorganized thinking. While delirium may affect any age group, it is most common in older patients, especially those with preexisting cognitive impairment. Patients with delirium after surgery recover more slowly than those without delirium and, as a result, have increased length of stay and hospital costs. The measured incidence of postoperative delirium varies with the type of surgery, the urgency of surgery, and the type and sensitivity of the delirium assessment. While generally considered a short-term condition, delirium can persist for months and is associated with poor cognitive and functional outcomes beyond the immediate postoperative period. In this article we will provide a guide to assess delirium risk preoperatively, and to prevent, diagnose, and treat this common and morbid condition. Care improvements such as identifying delirium risk preoperatively; training surgeons, anesthesiologists and nurses to screen for delirium; implementing delirium prevention programs; and developing standardized delirium treatment protocols may reduce the risk of delirium and its associated morbidity.
Delirium is an acute change in cognition characterized by inattention, fluctuating levels of consciousness, and/or disorganized thinking. Postoperative delirium is common, with significant associated morbidity and cost. Patients with delirium after surgery have high in-hospital mortality (4-17%)1-3 and 1-month4, 6-month5, 12-month6, and long-term mortality remains elevated.7,8 Additionally, delirium is associated with increased postoperative complications9, longer length of stay1, longer intensive care unit stay (ICU)3, and much higher rates of discharge to a nursing home.4,10 As a result, delirium adds significant cost to hospitalization and subsequent medical care.11
The incidence of postoperative delirium depends on the type of surgery. Table 1 highlights the incidence of delirium by surgical type. Hip fracture has the highest incidence of delirium, which is probably due to the urgent nature of the surgery and high comorbidity among these patients. Delirium is also common in patients after surgery for atherosclerosis pathology (cardiac, peripheral vascular, aneurysm repair). Elective and outpatient surgery have a lower, but still significant, incidence of delirium. A major factor in the variation of the reported incidence of delirium is the method for delirium assessment. For example in the cardiac surgery literature the methods used for delirium greatly influence the reported incidence of delirium: chart review only (3%)12, delirium noted during routine clinical care (8%)13, interviews with nurses (9%)7, and daily mental status testing and application of a validated diagnostic algorithm (53%).14
Because of the morbidity associated with delirium, all patients, especially older patients should be screened for delirium, at least, daily and more frequently if they are high risk. An algorithm for the diagnosis of delirium, the Confusion Assessment Method (CAM), which is based on Diagnostic and Statistical Manual of Mental Disorders (DSM)-IIIR criteria15, has been demonstrated to be reliable, sensitive, and specific for diagnosis of delirium compared to expert clinician examination.16,17 The algorithm for the CAM is displayed in Figure 1. Briefly, the criteria are a combination of Feature 1 (acute onset and fluctuating course), Feature 2 (inattention), and either Feature 3 (disorganized thinking) or Feature 4 (abnormal level of consciousness).
There are important elements of the CAM, which need to be clarified. First, attention is best assessed when formal testing (digit span, months of the year backward, serial 7s, etc) is combined with interviewer observations18. Importantly, orientation items have low sensitivity for inattention and delirium and should not be considered the standard assessment for attention.19 Additionally, there are two variants of delirium that are characterized as the hyperactive (agitated) variant and the hypoactive (quiet) variant of delirium. The hyperactive variant, which accounts for only about 25% of cases, is rarely missed, because the patient disrupts the flow of care20. The more common hypoactive variant is often missed because the patient is neither disruptive nor threatening21. For example, a patient with hypoactive delirium would briefly wake when addressed and may comply with some requests, but then quickly falls back to sleep. Several studies have found that the hypoactive variant is detected less frequently and carries a higher mortality, presumably from the delay in diagnosis.21-23 Thus, the CAM is a useful algorithm for diagnosis of delirium, but requires additional assessment of attention and observation.
The CAM-ICU operationalizes the CAM by adding objective assessments for attention, consciousness, and thought.24 The CAM-ICU has been validated in nonverbal ICU patients.24 The advantages of the CAM-ICU are that it can be performed by trained nurses or physicians; can be repeated over time to detect fluctuation and changes; and has been associated with ICU outcomes including mortality25, length of stay26, and cost27. The key elements of the CAM-ICU are the Richmond Agitation and Sedation Scale, a validated measure of consciousness28, the Attention Screening Exam29, and 5 thought questions. This information is used to complete the CAM algorithm for delirium.
Many preoperative risk factors for delirium have been described in the literature, and there are validated prediction rules for noncardiac and cardiac surgery to help identify those patients most at risk for postoperative delirium. The noncardiac surgery prediction rule identified 7 factors: age, impaired cognitive function, impaired physical function, abnormal laboratory values, alcohol abuse, thoracic surgery, and open-aortic surgery.2 The validated prediction rule for delirium after cardiac surgery identified 4 major risk factors: impaired cognitive function, low albumin, preoperative depressive symptoms, and prior stroke or transient ischemic attack.14 Table 2 describes the point scoring system for the prediction rules. In both rules, the incidence of delirium increases with increasing points so that the highest risk group is more likely to develop delirium than the lowest risk group (25× in the noncardiac surgery rule and 4× in the cardiac surgery rule). To complete the prediction rules, a thorough history and physical with screening for cognition, mood, and physical function are required. The paragraphs below describe risk factors for delirium in more detail.
The most common independent risk factor for delirium across studies is preexisting cognitive impairment. Preoperative cognitive screening is beneficial for assessing delirium risk, as well as, documenting baseline performance to detect delirium postoperatively. Cognitive screening should be performed with a standardized screening test. Many brief cognitive tests are available that require less than 10 minutes to complete.30,31 Importantly, orientation items and observation of standard conversation are not sufficient to assess for preoperative cognitive deficits.
Preoperative functional status is an independent risk factor for delirium after noncardiac surgery.2 The Activities of Daily Living (ADLs) and the Instrumental Activities of Daily Living (IADLs) provide an understanding of preoperative function. ADLs measure the ability to perform 7 basic care skills (feeding, bathing, grooming, using the toilet, transferring, and walking).32 The IADLs assesses the ability to perform 7 complex activities (using the telephone, grocery shopping, using transportation, cooking, housekeeping, taking medications, and handling finances).33 In addition to providing risk stratification for delirium, assessing this information preoperatively can inform the patient, the family, and the surgical team about the expected course of recovery postoperatively.
Abnormal preoperative laboratory values including glucose, sodium, potassium, and albumin are risk factors for delirium (Table 2).2,14 These abnormal laboratory values may represent underlying severe disease or organ system dysfunction, which is a predisposing risk factor for delirium. Hypoalbuminemia may be particularly important, because of its association with malnutrition, drug binding, fluid management34, and perioperative mortality.35 In the delirium prediction rule for medical patients, blood urea nitrogen to creatinine ratio ≥18, a marker of dehydration, was associated with incident delirium.36 Consistent with current practice, preoperative assessment of laboratory values can provide information about patients at high risk for delirium.
Many studies have identified depression as a risk factor for delirium after surgery.37-40 While the pathophysiology of this relationship remains to be determined, it is known that preoperative depression is associated with postoperative depression and incomplete recovery to independent functioning after surgery.41,42 The assessment of depression in older patients can be easily performed with the 15 question Geriatric Depression Scale, which assesses depressive symptoms using 15 yes / no questions43. The advantage of the Geriatric Depression Scale is that it can be self-completed by the older patient and scored by the clinician in a short time frame (3 minutes).44 In addition to delirium risk, assessment of depressive symptoms may provide insight into a patient's motivation for recovery.
Patients with multiple comorbidities are at increased risk of delirium. Alcohol abuse and prior stroke or transient ischemic attack deserve particular mention. Withdrawal from alcohol use has long been understood to precipitate delirium tremens, a variant of delirium.45 However, most postoperative delirium is not delirium tremens. Screening for alcohol use preoperatively can allow prevention of alcohol withdrawal with standardized protocols. Additionally, preexisting cerebral damage from prior strokes and transient ischemic attacks14, or a long history of alcohol abuse even in the absence of active drinking has been independently associated with postoperative delirium.2 Collection of this historical information from the patient or proxy before surgery is sufficient and routine cerebral imaging is not required for assessment of delirium risk.
In many older patients, the five senses decline with age. The combination of decreased sensory input (e.g., no glasses or hearing aids), cognitive impairment, and the perioperative environment may lead to misinterpretation of communication (i.e., talking back to television conversation), alarms (i.e., telephone ringing), and elements of the environment (i.e., window is a picture frame). Additionally, there is evidence in patients that increasing cognitive stimulation through improved sensory input may prevent delirium.46,47 As a result, preoperative assessment should include vision and hearing assessments and patients should be encouraged to bring their glasses and hearing amplifiers for use in the postoperative period to improve sensory input.
Delirium can be prevented in operative and medical patients46,47 by targeting moderate and high risk patients with clinical modules to improve baseline vulnerabilities and avoid iatrogenic complications. A summary of modules for prevention of postoperative delirium based on successful prevention models is presented in Table 3. For example, the nonpharmacological sleep protocols involved environmental changes conducive to sleep (i.e., lights off, create a relaxing environment, minimizing nighttime interruptions, dedicated time for sleep), which were successful in reducing psychoactive medication use and, ultimately, delirium47. Other key modules include improving sensory input, nutrition, ambulation, and preventing complications.
Haloperidol is a high-potency dopamine antagonist (antipsychotic) medication. In a single-site study, prophylactic administration of haloperidol did not reduce the incidence of delirium after hip fracture, but did reduce the severity and duration of delirium.48 However, severity scales tend to overweight hyperactive symptoms of delirium, so a possible explanation of these findings is the conversion of hyperactive delirium to hypoactive.49 As noted above, such a conversion may be a convenience for the care team, but actually worsens the prognosis of the patient. A follow-up study comparing high potency antipsychotic, atypical antipsychotic, and placebo in ICU patients found no difference in the days alive without delirium.50 As a result, the practice of prophylaxis with antipsychotics should be avoided at present due to increased risk of death, delirium, and complications in older patients attributed to this class of drugs.51
Acetylcholinesterase inhibitors are medications used to stabilize cognitive function in Alzheimer Disease. A randomized controlled trial of rivastigmine for delirium prevention in cardiac surgery found no effect of treatment on delirium incidence or cognitive performance.52 In elective orthopedic surgery, the results have been mixed, with one trial showing no effect53 and another suggesting benefit54. A randomized controlled trial of rivastigmine for delirium treatment was stopped early due to increased mortality in the treatment arm and no effect on delirium.55 Thus, at this time, prevention or treatment of delirium with acetylcholinesterase inhibitors should be avoided.56
Dexmedetomidine is an alpha-2 adrengergic receptor agonist that is used for sedation. Two recent studies have found that the use of dexmedetomidine for sedation in the ICU setting has a reduced rate of delirium compared to midazolam and lorazepam.57,58 Additionally, a randomized trial of intraoperative sedation with dexmedetomidine, propofol, or midazolam found that dexmetometadine was associated with a lower incidence of postoperative delirium.59 Thus, the use of dexmedetomidine in patients at intermediate-risk and high-risk for delirium may have benefits that outweigh the potential adverse events.
The delirium risk factors described above are classified as predisposing factors, that is, factors that can be assessed before surgery and increase the patient's risk. Precipitating factors, a distinct class of risk factors, occur intraoperatively and postoperatively and are thought to acutely precipitate the delirium episode. Precipitating factors for delirium have been more difficult to identify than predisposing factors. A primary challenge has been heterogeneity, due to patient factors (age, education, comorbidity), surgery factors (type of surgery, techniques used, hypothermia, bleeding), physiologic factors (inflammation, microembolization, blood-brain barrier function), intraoperative factors (anesthesia, cerebral oxygenation, hypotension), perioperative factors (medication, sleep, complications), and postoperative factors (rehabilitation, depression, social supports). Thus, all of the precipitating factors for postoperative delirium have not been fully elucidated. Table 4 summarizes predisposing (preoperative), and known precipitating (intraoperative, and postoperative) factors that may be associated with delirium.
During surgery, numerous medications with cognitive properties are given to patients. Inhaled anesthetics alter electrical activity in the brain60 and have been associated with amyloid deposition and apoptosis61. Induction drugs and benzodiazepines have significant cognitive properties which may precipitate delirium62. While regional anesthesia has the potential to reduce this exposure, studies of general versus regional anesthesia have not demonstrated a reduction in delirium.63 One reason for this may be the concomitant administration of sedatives in addition to the regional anesthetic. In a recent randomized study, lighter depth of sedation, measured with the Bispectral Index, resulted in 50% less postoperative delirium than deep sedation.64
Pain medications may precipitate delirium, particularily meperidine, which increases the odds of delirium over other opioids62. Many of these medications are necessary for the operation; however, the clinically important point is to recognize the risk for delirium associated with cognitively active medications and minimize the exposure to these medications, especially in older patients with cognitive impairment.
After surgery, many patients are given medications that can impair cognitive function. For example, in the postoperative tracheally intubated patient, sedatives such as benzodiazepines or propofol are given. In these patients, dexmedetomidine used for sedation may reduce the risk of delirium59. Out of the ICU setting, benzodiazepine and sedative use may be minimized by improving sleep hygiene using nonpharmacological measures such as decreasing environmental noise, creating a relaxing environment, and preserving the circadian rhythm47.
While opioids may precipitate delirium, uncontrolled pain may also precipitate delirium.62 When opioid pain medication is needed after surgery, standardized age-adjusted protocols should be used to treat pain and taper opioid doses. Strong consideration should be given to standing pain medication, especially acetaminophen, which has been shown to reduce total opioid needs and improve patient reports of pain in a postoperative randomized controlled trial65. The advantage of acetaminophen is its limited cognitive properties, compared to opioids. Even in patients who require opioids, administration on a schedule has been shown to reduce total dose needs relative to as needed (PRN) dosing.66 Patient-controlled analgesia improves pain control, however, caution should be used in older patients with cognitive impairments and those who have developed delirium.67 In summary, postoperative sedative use should be minimized, and postoperative analgesia should be administered using rational, carefully designed protocols designed to minimize systemic exposure to opioids with psychoactive properties.
After cardiac surgery, patients are transferred to the ICU environment, which is busy, noisy, and light-filled where patients are approached, assessed, and stimulated constantly. Recent work in the ICU setting found that the environment may contribute to delirium, through sleep deprivation and overstimulation.68 While this environment may be required in the immediate postoperative period, early transfer of medically stable patients to less intense wards should be considered. Additionally, consideration should be given to balance the patient's monitoring needs with the sleep requirements of the patient (i.e., Does the patient need a standing order for vital signs at midnight and 4 a.m.?). Even the non-ICU environment can be disorienting: preservation of the sleep-wake cycle (minimized nighttime interruptions, adequate lighting, sleep hygiene), provision of orienting supplies (clock, calendar, orientation board), and cognitively stimulating activities (glasses, hearing amplifiers, puzzles) may minimize precipitation of delirium.46,47
Complications of hospitalization and surgery can precipitate delirium. For example, a leading identifiable cause of delirium in older inpatients is urinary tract infection associated with catheter use69. Preventable medical processes such as deep venous thrombosis, pressure ulcer, deconditioning, malnutrition, and dehydration should be assessed using a standardized team-based approach46. Additionally, reduced mobility through formal restraints or informal tethers (i.e., IV lines, oxygen tubing, urinary catheters, etc) can contribute to delirium, loss of function, falls, and increased rehabilitation placement.70 The routine use of care systems to prevent postoperative complications may also prevent delirium.
The primary treatment of delirium is to identify and treat its underlying causes. Thus, the clinician is recommended to begin with a broad differential diagnosis and systematically eliminate potential causes. It should be noted that delirium is associated with significant morbidity and mortality and thus, all patients with delirium should be assessed promptly with an interim history, thorough physical examination with a focus on the neurological examination, and targeted laboratory testing based on the history and examination. Most delirious patients require at least basic laboratory including a complete blood count, basic metabolic panel including renal function, and urinalysis. It is also important to carefully review the patient's medications, particularly the nursing administration record where it is clear exactly what medications the patient received and when. It is also important to note that the causes of delirium are often multifactorial, and therefore the search for additional causes and contributors to delirium should not be terminated when a single cause has been identified.
Prior work has found that in the absence of focal neurological deficits, a head computed tomography has low diagnostic value in the assessment and treatment of the delirious patient.71 MRI scanning in the postoperative patient is difficult, because of the acuity of illness, recently implanted hardware (e.g., staples prostheses, grafts, valves, etc), and the time and cooperation required for scanning. In the patient with delirium, sedation, which can worsen or prolong the delirium, may be needed for imaging72. In the patient after cardiac surgery, there will likely be new foci on imaging related to microemboli during the operation.73 Because the causal link between microemboli and delirium has not been established74, the clinical significance of such findings is unknown. Additionally, stroke thrombolysis is contraindicated in the postoperative surgery patient75, so the potential treatments for stroke are limited in this setting. The treatment of occlusive infarct in postoperative patients is identical to the treatment of cardiac disease (e.g., aspirin, statin, arterial blood pressure control, cardiac risk reduction, rehabilitation). Therefore, cerebral imaging should be restricted to those with new focal neurological findings, or those at very high risk in whom no other cause of delirium can be identified.
For patients who develop agitation, a through review of the medications and physical examination, including pain assessment, is required. First, the offending precipitant of the delirium (constipation, urinary retention, etc.) should be relieved. Nonpharmacologic treatments should also be initiated, regardless if the cause is identified. For example, elimination of environmental noise, allowing the patient to sleep at night, and reorientation efforts should be implemented. A model of care for the delirious patient found that environmental modifications and staff training could produce reductions in patient agitation, reduction in use of psychoactive medications, with similar length of stay.76 Another useful resource is family members who can serve as a reorienting and reassuring stimulus. Because of the low risk of adverse events, nonpharmacological methods are recommended as a first step.
For patients in whom these nonpharmacological interventions are not sufficient, antipsychotics are considered the first line for the pharmacological management of agitation associated with delirium.51,77 For most patients, haloperidol at a low initial dosage of 0.5 to 1.0mg is a reasonable choice. If there is no response within one hour, a repeat dosage may be considered. If there is no effect after 2-3mg of haloperidol, it is unlikely that the patient is going to respond. Higher doses administered IV are frequently used in ICU patients.50 Antipsychotics administered in the acute setting have not been demonstrated to have increased mortality, but even intermediate-term (6-12 weeks) use of antipsychotics is considered to carry an increased mortality, especially in cognitively impaired patients.78,79 Additionally, electrocardiograms should be performed at baseline and monitored in high-risk patients (i.e., older, atherosclerosis pathology, major surgery) taking high-dose antipsychotics due the risk of QTc prolongation.80 Finally, early evidence suggests that acute administration of antipsychotics may be associated with oropharyngeal dysphagia which may further delay recovery81. At this time, there is no evidence for a incremental benefit of atypical antipsychotics beyond that of haloperidol for the treatment of delirium.51
For patients with contraindications to antipsychotics such as Parkinson's disease, Lewy Body dementia, prior seizures, and prior neuroleptic malignant syndrome; agitation may be better managed with benzodiazepines. In general, benzodiazepines disinhibit patients and patients should be monitored for a paradoxical reaction, where administration of the benzodiazepine results in agitation. Additionally, prior work has shown that benzodiazepines may actually prolong or worsen the course of delirium72. Finally, respiratory depression becomes a risk in older patients with respiratory comorbidities who have just undergone surgery. Thus, other than the specific case of alcohol or chronic sedative withdrawal, use of benzodiazepines should not be considered a first-line therapy and should be reserved for cases in which clinical circumstances limit use of antipsychotics.
While generally thought of as a short-term disorder, delirium can have lasting effects beyond the perioperative period. First, delirium itself can persist for months. In a study of patients with delirium upon admission to a rehabilitation facility after hospitalization, delirium persisted for 6 months in one-third of patients.82 Persistent delirium increased the 1-year mortality and prevented functional recovery. 82,83 There is an increasing body of evidence that persistent delirium can delay both cognitive and functional recovery.10
Delirium may accelerate the cognitive decline in patients with Alzheimer's Disease84. While delirium is felt to be distinct from postoperative cognitive dysfunction, the two syndromes may be highly correlated in the short-term (1 week) 85-87. The impact of delirium on long-term postoperative cognitive dysfunction has been less consistent.85,86,88 In addition to cognitive function, delirium has been associated with postoperative depression,37,38 another factor which may impede recovery. Finally, newer evidence is emerging that (younger) patients with delirium may develop a posttraumatic stress disorder-like syndrome.89,90 As a result, delirium may have long-term mental health complications that are not fully studied and may impact on functional recovery.
Several studies have demonstrated that postoperative delirium is associated with functional decline and nursing home placement 1-3 months after surgery.4,91 However, these same studies have not demonstrated a consistent association between delirium and functional decline six months or longer after surgery. This difference may be related to reduced statistical power due to decreased functional decline rates in the long term and dropout of the patients with delirium.4,91 Future work in larger epidemiological studies will help determine the relationship between the incidence of delirium and long-term functional change after surgery.
Delirium is an acute change in cognitive function, specifically attention, with associated disorganization of thought and abnormal level of consciousness. Postoperative delirium is very common, especially in older surgical patients, and is associated with substantial morbidity, costs, and mortality. Preoperative delirium risk assessment is critical for identification of those patients who would most benefit from delirium prevention and surveillance protocols. Nonpharmacologic delirium prevention strategies have proven effective at reducing delirium incidence, but pharmacological prevention strategies do not yet have trial-based support. The primary treatment of delirium is to identify and treat the underlying causes. Delirium has substantial long-term consequences, which are currently being better defined through large-scale epidemiological studies. Assessing preoperative delirium risk, using delirium prevention strategies, and implementing standardized treatment protocols are important components of optimal care for older patients undergoing surgery.
Funding: Dr. Rudolph is the recipient of a VA Rehabilitation Research Career Development Award and Dr. Marcantonio is a recipient of a Mid-Career Investigator Award in Patient-oriented Research from the National Institute on Aging (K24 AG035075).
The authors declare no conflicts of interest.
DisclosuresName: James L. Rudolph, MD, SM
Contribution: Study design, conduct of study, data collection, data analysis and manuscript preparation.
Name: Edward R. Marcantonio, MD, SM
Contribution: Study design, conduct of study, data collection, data analysis and manuscript preparation.
James L. Rudolph, Geriatric Research, Education, and Clinical Center, VA Boston Healthcare System, Boston, MA, USA; Division of Aging, Brigham and Women's Hospital, Boston, MA, USA; Harvard Medical School, Boston, MA.
Edward R. Marcantonio, Divisions of General Medicine and Primary Care and Gerontology, Beth Israel Deaconess Medical Center, Boston, MA, USA; Harvard Medical School, Boston, MA.