POSTOPERATIVE DELIRIUM

Abstract

Delirium is an unfortunately common complication seen during the postoperative course. Because of its significant association with physical and cognitive morbidity, clinicians should be aware of evidence-based practices relating to the diagnosis, treatment, and prevention of postoperative delirium. Here, we review selected recent literature pertaining to the epidemiology and impact of the condition, perioperative risk factors for its development and/or exacerbation, and strategies for management of delirium, with additional attention to the intensive care unit population.

Introduction

According to the Roman philologist Marcus Terentius Varro (116 B.C. -27 B.C.) in De lingua Latina, “delirium” is a term of Latin origin coming from the agricultural activity of ploughing¹. The term literally means “going off the ploughed track”; its figurative meaning is similar to the concept of “becoming mad”. It was likely used for the first time in medical language by Aulus Cornelius Celsus (25 B.C.- 50 A.C.) in his work De Medicina: “Now it is useless to adopt remedies when the delirium is at its height… there is nothing else to do than to restrain the patient, but when circumstances permit, relief must be given with haste…”².

The current DSM-IV definition of delirium describes an acute and fluctuating disturbance of consciousness with reduced ability to focus, maintain, or shift attention, accompanied by change in cognition and perceptual disturbances secondary to a general medical condition. In neo-Latin languages, that same condition is identified by expressions like acute confusional state (“stato confusionale acuto” in Italian, for example)³. The definition of delirium has been further refined by the International Classification of Diseases 10^th edition (ICD-10) to include psychomotor disturbances, disturbance of the sleep-wake cycle, and emotional disturbances. In most cases delirium has a transient and self-limiting nature (most cases recover within 4 weeks), but in up to 50% of cases delirium can persist up to the moment of hospital discharge⁴.

The etiology of delirium is not yet fully understood and is probably multifactorial. Acute central cholinergic deficiency is one of the most widely-accepted explanatory theories; decreased GABA-ergic activity, abnormalities in melatonin and serotonin pathways, noradrenergic hyperactivity, inflammation with increase release of IL-1 and IL-6 leading to neuronal damage, and cerebral hypoperfusion are other possible pathogenetic mechanisms. The current understanding of the neurobiology of delirium is thoughtfully addressed in two recent reviews^{5, 6}.

Post-operative delirium (POD) is a form of delirium that manifests in patients who have undergone surgical procedures and anesthesia, usually peaking between one and three days after their operation. It must be differentiated from emergence delirium, which occurs in 8 to 20% of patients after awakening from general anesthesia, especially in younger ages^{7, 8}. The causes of and the possibility of a relationship among emergence delirium, POD, postoperative cognitive decline, and postoperative incident dementia have not yet been fully elucidated⁹.

Epidemiology

Delirium is very common in hospitalized patients; its prevalence in the adult general medicine population is 10–24%, and it affects 37–46% of the general surgical population. In the intensive care unit (ICU) setting, delirium has been reported in up to 87% of patients. Postoperative delirium rates vary widely, ranging from 9 to 87% depending on the age of patients and the type of surgery. For a thorough review of prevalence rates in various medical settings, please refer to Maldonado’s excellent review¹⁰. The degree of operative stress appears to be related to the prevalence of post-operative delirium. Furthermore, low pre-operative executive scores and depression independently predict delirium in older individuals¹¹.

Implications and Consequences

Delirium is associated with increased days of mechanical ventilation and ICU length of stay¹², increased hospital length of stay¹³, and patients’ functional decline¹⁴. Furthermore, all-cause mortality increases by at least 10–20% for every 48 hours of delirium^{13, 15, 16}. A recent meta-analysis demonstrated that delirium during hospitalization doubles a patient’s risk of post-discharge institutionalization and death, and increases risk of dementia up to 10-fold. Impressively, these poor outcomes were independent of common confounders like age, sex, and comorbidities¹⁷. Additionally, in-hospital delirium is associated with more than doubled one-year medical costs^{18, 19}. Some of the cost and mortality implications arise from the association of delirium with other postoperative complications, including falls, pressure ulcers, urinary tract infection, respiratory difficulties, myocardial infarction, and atrial fibrillation^{11, 19}.

Diagnosis

As it is increasingly evident that delirium is associated with poor outcome and increased health care costs, its timely diagnosis is crucial to prevent patients from developing severe long-lasting complications. Only a fraction of patients with delirium are quickly recognized by care-givers. Those patients usually present with the hyperactive form of delirium, commonly characterized by restlessness, agitation, hallucinations and delusions. In contrast, the hypoactive form of delirium, characterized by reduced movements, paucity of speech, and unresponsiveness, can be misdiagnosed as depression, anxiety, or even as a calm and comfortable patient when proper screening is not performed. A rough distribution of delirium presentations suggest that it may be of the hyperactive form in 25%, hypoactive in 50%, and mixed in 25% of cases^{20, 21}. The more clinically “silent” hypoactive type may be associated with greater mortality than hyperactive types²⁰.

As missing a diagnosis of delirium might have important negative consequences on outcome, clinicians should actively look for it, especially in high risk patients (Table 1). Screening must be performed in all vulnerable patients to minimize the impact of this condition, whether the diagnosis is suspected on clinical grounds (e.g. in the agitated patient) or not (as is common in hypoactive delirium). The gold standard is a psychiatric evaluation according to DSM criteria. Appearance, level of consciousness, thought and attention, speech, orientation, memory, mood, judgment, and behavior should all be assessed²². Examiners without psychiatric training can effectively assess patients using one of the validated screening and diagnostic instruments. A recent systematic review²³ found that current evidence supports the use of the Confusion Assessment Method (CAM)²⁴. For intubated patients and in ICU populations, the CAM-ICU is a good alternative choice²⁵. It is important to note that the altered cognition and inattention of delirium are sometimes impossible to differentiate from those seen in dementia; they must be distinguished on the basis of time course (acute for delirium, chronic in dementia).

Table I.

Risk factors for delirium.

Predisposing19	Precipitating19, 21
Reduced cognitive reserve:	Medications or medication withdrawal:
Dementia	Anticholinergics
Depression	Muscle relaxants
Advanced age	Antihistamines
Reduced physical reserve:	GI antispasmodics
Atherosclerotic disease	Opioid analgesics29
Renal impairment	Antiarrhythmics
Pulmonary disease	Corticosteroids
Advanced age	>6 total medications
Preoperative beta blockade30	>3 new inpatient medications
Sensory impairment (vision, hearing)	Pain31
Alcohol abuse	Hypoxemia
Malnutrition	Electrolyte abnormalities
Dehydration29	Malnutrition
Apolipoprotein E4 genotype32	Dehydration29
	Environmental change (e.g. ICU admission)
	Sleep-wake cycle disturbances33
	Urinary catheter use
	Restraint use
	Infection
	Psychotropic medications:
	Antidepressants
	Antiepileptics
	Antipsychotics
	Benzodiazepines

In diagnosing delirium, there are no definitive laboratory tests, and the differential diagnosis for acute abnormalities of cognition and attention is broad. A complete blood count, electrolyte panel, glucose measurement, arterial blood gas, urine analysis, and electrocardiography should be considered in all patients with acute confusion to rule out correctable conditions like hypoxia, hypoglycemia, and electrolyte imbalance. Life-threatening organic conditions including withdrawal syndromes (e.g. ethanol) and intracranial processes such as meningitis, hypertensive crisis, and status epilepticus also may explain an acute change in consciousness, and should be given appropriate consideration. Imaging studies, including chest radiography and computed tomography or magnetic resonance imaging of the head, should only be requested based on clinical suspicion of specific pathologies such as stroke. There are currently no known neuro-imaging correlates of delirium. Electroencephalography may show non-specific changes, particularly generalized slowing to the theta-delta range²⁶.

Risk Factors: Predisposing and Precipitating Factors

Postoperative delirium is not preventable in all patients, but it may be prevented in some, and its severity and duration may be limited in others, especially if promptly recognized and managed²⁷.

According to Inouye, delirium development follows a pathogenetic model in which predisposing and precipitating factors can be identified. Predisposing factors are intrinsic to patients, are already present in the preoperative period, and determine vulnerability to brain insults (Table I). The prevailing theory is that patients prone to delirium have a diminished brain reserve; in them, delirium may occur with noxious insults that would not cause major cognitive change in subjects with otherwise intact cognitive reserve²⁸. For instance, dementia consistently emerges as one of the most significant risk factors for delirium. Precipitating factors are potentially modifiable factors associated with hospitalization, particularly in the postoperative period (Table I).

Intraoperative Risk Factors for Delirium

The relationship between anesthesia and delirium is complex, and not yet fully elucidated. On one side, we know that delirium is common after anesthesia and surgery. Delirium is certainly associated with many classes of drugs used in the perioperative period (Table I). The Beers criteria for inappropriate medication use in the elderly were developed in 1997 and updated in 2003³⁴. Although the guidelines were developed for use in the outpatient setting, many of the cautions are appropriately considered in the perioperative setting as well. Potentially inappropriate classes of medications for use in older adults that theoretically increase risk for postoperative delirium include antihistamines, benzodiazepines, muscle relaxants, and the analgesic meperidine, which is often used for postoperative shivering³⁴. Those medications should be avoided in vulnerable patients as much as possible.

Perioperative patients are also exposed to many precipitating factors that may lead to delirium, particularly pain, environmental unfamiliarity including disrupted sleep cycles, and use of a urinary catheter (Table I). Clinicians must take care in treating the former two precipitants, as certain classes of analgesics and sedatives will perpetuate or exacerbate existing or developing delirium. Current recommendations for the treatment of pain in cognitively vulnerable patients are addressed in a recent review article³¹.

Vascular and cardiac surgeries are associated with a high incidence of delirium, possibly because cerebral complications of atherosclerosis may also reduce brain reserve. Orthopedic surgery, especially urgent or semi-urgent repair of hip fractures, is another clinical situation where high incidence of delirium is seen. The strategy of choosing regional anesthesia, in lieu of general anesthesia, for orthopedic procedures including hip replacements has recently proven successful in reducing the incidence of postoperative delirium in elderly patients³⁵. A smaller prior study showed no difference in delirium rates but more frequent impairment in postoperative cognition following general anesthesia compared with a regional technique³⁶.

Prevention of Postoperative Delirium

The astute clinician may be presented with many opportunities to decrease a patient’s risk of experiencing delirium during the postoperative period (Table II). While this article focuses on pharmacological strategies, it is essential to note that protocols of behavioral intervention^{27, 37} have been shown to have a substantial effect on delirium risk in the elderly. Education of physicians and support staff and implementation of simple additional care measures, including deliberately orienting the patient to place, time and reason for hospitalization, encouraging early mobilization, and reminding patients to use their own glasses and hearing aids when appropriate, not only reduces delirium risk (odds ratio 0.4, 95% confidence interval [CI] 0.24–0.77) but also prevents functional decline in some elderly patients, while not requiring additional staff for implementation³⁷. These measures must be implemented as institutional standard of care whenever possible, and particularly in the ICU setting.

Table II.

Perioperative interventions for delirium.

	Perioperative interventions
	Intraoperative prevention	Postoperative prevention	Treatment	ICU sedation
Helpful/Recommended		Medication awareness34 Behavioral 27, 37	Antipsychotics38, 39 Behavioral 27, 37
Not recommended/Harmful			BZP40 Rivastigmine41–43	BZP44
Unclear	Low-dose ketamine on induction45 Regional anesthesia with or without sedation35, 36	Antipsychotics46, 47 AChEase inhib48–50 Statins51, 52		Dexmed53

Abbreviations: Dexmed, dexmedetomidine. BZP, benzodiazepines. AChEase inhib, acetylcholinesterase inhibitors

Emergence from operative anesthesia

The phenomenon of emergence delirium is well described in the pediatric population but infrequently addressed in adults. In a recent issue of this Journal, Radtke and colleagues found hyperactive Richmond Agitation and Sedation Scale (RASS) values, consistent with the common understanding of emergence delirium, in approximately 5% of adult patients in the post-anesthesia care unit (PACU)⁷; a study by Yu and colleagues found an incidence of closer to 20%⁸. Further work is essential to define the relationship of emergence delirium to postoperative delirium in its more formal definitions, and to address the impact of emergence delirium on postoperative outcomes.

Prevention of delirium in postoperative patients

Consistent with its apparently multifactorial etiology, pharmacological strategies for prevention address many different pathways of pathology thought to be involved in delirium.

Antiinflammatory drugs such as the HMG-CoA reductase inhibitors, or “statins”, and ketamine have been the subject of recent investigation. After a massive epidemiological study demonstrated significantly increased rates of documented ICD-9 codes for delirium in elderly patients undergoing elective surgery who were taking statins⁵², a second study demonstrated a significant reduction with statins, with an odds ratio of 0.54 (95% CI 0.35–0.84)⁵¹. The first study was hampered with significant methodological concerns including the lack of sensitivity of ICD-9 codes as a surrogate for this diagnosis; the second incorporated only 122 cases of delirium. A final answer on this important question awaits a prospective randomized controlled trial. A small prospective randomized controlled trial on delirium following cardiac surgery with cardiopulmonary bypass demonstrated that a single intravenous bolus of 0.5 mg/kg ketamine during anesthetic induction was associated with decreased postoperative delirium and a reduction in the inflammatory marker C-reactive protein⁴⁵.

Modulation of dopaminergic transmission figures prominently in the prophylaxis and treatment of delirium. Prophylactic low-dose haloperidol, olanzapine, or risperidone may be effective in decreasing both the severity and duration of delirium episodes, though they do not prevent development of delirium itself³⁹. A trial comparing haloperidol, ziprasidone and placebo for prevention of delirium found no difference in the primary outcome of days alive without delirium or coma among the three groups⁴⁶. A more promising result was achieved in a small randomized trial, where a single sublingual dose of risperidone versus placebo immediately upon return to consciousness impressively attenuated the risk of post-cardiac surgery delirium⁴⁷. Importantly, in no trial has the administration of an antipsychotic drug been associated with side effects, including extrapyramidal symptoms, although the risk of harm to vulnerable patients (in particular, those with Parkinson’s disease or dementia with Lewy bodies) should not be underestimated. All trials have been small, and while the evidence is suggestive, further investigation is needed.

Despite the clear importance of cholinergic transmission in the pathogenesis of delirium, the role of acetylcholinesterase inhibitors in prevention of postoperative delirium has not been demonstrated. Two small randomized, placebo-controlled trials of donepezil failed to reduce incident delirium^{49, 50}; a randomized double-blind placebo-controlled trial of rivastigmine, which enrolled only 120 patients, also found no difference in delirium or other cognitive outcomes⁴⁸.

Sedation in the ICU

ICU patients deserve special consideration. The use of sedation for patient safety and comfort (recently reviewed by Morandi and colleagues⁵⁴), particularly in those receiving mechanical ventilation, is a particularly germane intervention from the standpoint of delirium. With several pharmacologic agents available for ICU sedation, some have been shown to be more appropriate for use in patients with high risk of delirium. In general, unnecessary polypharmacy must be avoided in these already complex patients; in other words, sedation should only be used if clinically indicated, and at the lowest tolerated dose. There is a growing consensus that, in the absence of other clear indications such as benzodiazepine or alcohol withdrawal delirium, benzodiazepines are not agents of choice for ICU sedation⁴⁴.

Several small randomized studies published in high-impact journals have suggested that the alpha-2 agonist dexmedetomidine may decrease the incidence of delirium; however, a recent meta-analysis concluded that it confers no significant benefit (relative risk for delirium with dexmedetomidine versus other sedative regimens = 0.85 [95% CI 0.64–1.13]), although the confidence interval remains wide⁵³. Of interest, a small randomized trial found that an infusion of the alpha-2 agonist clonidine reduced the incidence of delirium during weaning from mechanical ventilation in the intensive care unit⁵⁵. As dexmedetomidine has not earned the approval of the European Medicines Agency for human use, the authors of that study suggested clonidine would be a worthwhile subject of further research into prevention of delirium in the ICU setting.

Treatment of Postoperative Delirium

Should delirium be detected in a patient, either postoperatively or in the ICU setting, some treatment strategies are available and supported by the peer-reviewed literature. Unsurprisingly, many of the therapies attempted for prevention have been successful in the treatment of established delirium. It bears repeating that there is seldom a contraindication to implementing and/or emphasizing the recommended behavioral interventions described above for delirium prevention as a treatment strategy in an already delirious patient, although they have not been formally evaluated in that context³⁷.

The best-studied classes of pharmacological treatments for delirium are the typical and atypical antipsychotics. The first-generation antipsychotics haloperidol and chlorpromazine, as well as the second-generation antipsychotics olanzapine and risperidone, all appear to be equally effective treatments for established delirium, with the caveat that studies in this area tended to be small, and single-center, with limited generalizability (for review, see the 2007 Cochrane review³⁹). The first published multicenter, randomized trial comparing an antipsychotic – in this case, quetiapine – with placebo for treatment of incident delirium found significant reductions in time spent delirious and time spent agitated for the treatment group³⁸. The study was limited by its small size, and interpretation is somewhat difficult since open-label haloperidol could be used at the physician’s discretion.

As with prevention of delirium, acetylcholinesterase inhibitors have been investigated in its treatment. After two very small trials of rivastigmine – one nonrandomized and open-label⁴¹, one blinded and placebo-controlled⁴² – suggested a beneficial effect, a larger double-blind, placebo-controlled study was undertaken to enroll 440 patients in 6 hospitals. The study was halted after a total of 104 patients completed the trial, when monitoring indicated that the death rate in the treatment group was 12 of 54, compared to 4 of 50 patients taking placebo⁴³. The available evidence does not support the use of rivastigmine in these patients. Other acetylcholinesterase inhibitors, including donepezil and galantamine, are even less studied. Benzodiazepines are implicated in the development of delirium, and there is no convincing evidence to support their use in its treatment⁴⁰.

Delirium, Postoperative Cognitive Decline, and Postoperative Incident Dementia

Evidence is emerging that people who have delirium are at increased risk for persistent cognitive decline, and even for progression of dementia^{12, 14, 17, 20, 56–58}. A recent study examined the impact of delirium on the trajectory of cognitive function in a cohort of patients with Alzheimer’s disease, and found significant acceleration of cognitive decline following an episode of delirium⁵⁶. For older patients with and without dementia, delirium is an independent predictor of sustained poor cognitive and functional status during the year after surgery¹⁴. In an important recent study, elderly people who were hospitalized for surgical or medical illness had a 40% increased risk of incident dementia compared with a matched cohort who were not hospitalized (adjusted hazard ratio 1.4 [95% CI, 1.1 to 1.7])⁵⁷. A chart review of elderly patients admitted for hip or knee replacement demonstrated a strong relationship between the occurrence of postoperative delirium and long-term risk of dementia⁵⁸.

Surgery is characterized by a marked inflammatory response, and inflammation has been implicated both in the pathogenesis of delirium, and, in the case of persistent inflammation, in the pathogenesis of cognitive decline in Alzheimer’s disease⁵⁹. The available evidence demonstrates that surgery and anesthesia, and indeed many major health stressors, are associated with delirium, and development of delirium is associated with later incident dementia. It is unknown whether surgery and anesthesia contribute to dementia risk independent of other comorbid features of ill health. It is, however, conceivable that there exist vulnerable subgroups, based on such factors as genetic susceptibility or co-morbidities, in whom exposure to surgery or anesthesia unmasks the premorbid decline in cognitive reserve associated with later progression to frank dementia.

Conclusion

Postoperative delirium is common, under-diagnosed by healthcare practitioners, potentially preventable in some instances, and associated with increased morbidity and resource utilization. In the non-surgical setting it has been established that delirium is associated with dementia and is a harbinger of earlier death. The relationships among postoperative delirium and subsequent cognitive decline, incident dementia and mortality have not been well characterized and should be explored rigorously in future prospective trials. It is essential for patient care that strategies to prevent, diagnose and treat postoperative delirium should be prioritized by all practitioners in the perioperative period.