Neuropathogenesis of Delirium: Reviewof Current Etiologic Theories

and Common Pathways

large neutral amino acids, neuroendocrine, oxidative stresshttp://dx.doi.org/10.1016/j.jagp.2013.09.005Received May 23, 2012; revised September 10, 2013; accepted September 13, 2013. From the Departments of Psychiatry, Internal Medicine &Surgery and the Psychosomatic Medicine Service, Stanford University School of Medicine, and Board of Directors, American DeliriumSociety, Stanford, CA. Send correspondence and reprint requests to Jose R. Maldonado, M.D., F.A.P.M., F.A.C.F.E., Departments ofPsychiatry, Internal Medicine & Surgery, and the Psychosomatic Medicine Service, Stanford University School of Medicine, 401 Quarry Rd.,Ofce #2317, Stanford, CA 94305. e-mail: jrm@stanford.edu

2013 American Association for Geriatric Psychiatry1190Jose R. Maldonado, M.D., F.A.P.M., F.A.C.F.E.

Delirium is a neurobehavioral syndrome caused by dysregulation of neuronalactivity secondary to systemic disturbances. Over time, a number of theories havebeen proposed in an attempt to explain the processes leading to the development ofdelirium. Each proposed theory has focused on a specic mechanism or pathologicprocess (e.g., dopamine excess or acetylcholine deciency theories), observational andexperiential evidence (e.g., sleep deprivation, aging), or empirical data (e.g., specicpharmacologic agents association with postoperative delirium, intraoperativehypoxia). This article represents a review of published literature and summarizes thetop seven proposed theories and their interrelation. This review includes the neu-roinammatory, neuronal aging, oxidative stress, neurotransmitter deciency,neuroendocrine, diurnal dysregulation, and network disconnectivity hypothe-ses. Most of these theories are complementary, rather than competing, with manyareas of intersection and reciprocal inuence. The literature suggests that manyfactors or mechanisms included in these theories lead to a nal common outcomeassociated with an alteration in neurotransmitter synthesis, function, and/or avail-ability that mediates the complex behavioral and cognitive changes observed indelirium. In general, the most commonly described neurotransmitter changes asso-ciated with delirium include deciencies in acetylcholine and/or melatonin avail-ability; excess in dopamine, norepinephrine, and/or glutamate release; and variablealterations (e.g., either a decreased or increased activity, depending on deliriumpresentation and cause) in serotonin, histamine, and/or g-aminobutyric acid. In theend, it is unlikely that any one of these theories is fully capable of explaining theetiology or phenomenologic manifestations of delirium but rather that two or moreof these, if not all, act together to lead to the biochemical derangement and, ulti-mately, to the complex cognitive and behavioral changes characteristic of delirium.(Am J Geriatr Psychiatry 2013; 21:1190e1222)

Key Words: Aging, delirium, encephalopathy, physiologic stress, neuroinammation,Am J Geriatr Psychiatry 21:12, December 2013

attention) and awareness (reduced orientation to theenvironment), cognition (e.g., memory decit, disori-

developed in-hospital delirium versus 51% for con-trol patients (risk ratio: 2.24), even after accounting

Maldonadoentation), language, visuospatial ability, or perceptionthat is not better explained by a preexisting, estab-lished, or other evolving neurocognitive disorder.2e5

A recent study at a large general hospital found thatthe point prevalence of delirium is one in every veinpatients, with some variation in certain medicosur-gical units and populations.6 Studies have found thatbetween 14% and 24% of elderly patients are admittedto the hospital with delirium7,8 and that deliriumoccurs in up to 50% of elderly inpatients.9 In theemergency department, the prevalence in elderlypatients is 9.6%.10 The reported rate of postoperativedelirium ranges between 10% and 74% depending onthe type of surgery and population under study.11e15

Studies have demonstrated that up to 87% of criticallyill patients develop delirium during their intensivecare unit (ICU) stay.16 This is important consideringthat the proportion of patients in the ICU aged at least65 years is 56%.17 Furthermore, studies predict that by2015 the rate of elderly aged 80 years and olderadmitted to the ICUwill increase by 72%, representingroughly one in four admissions to the ICU.18

DELIRIUM CONSEQUENCES

Delirium has been reported to be one of the sixleading causes of preventable conditions in hospi-talized elderly patients.19 After controlling fordemographics, apparent illness severity, age, andmedical comorbidities, patients who developdelirium fare much worse than their nondeliriouscounterparts. The mortality rate for elderly patientsin acute care hospitals is much higher among thosewith delirium than those without delirium: 8% versusDELIRIUM: SCOPE OF THE PROBLEM

Delirium is a neurobehavioral syndrome caused bydysregulation of baseline neuronal activity secondaryto systemic disturbances,1 characterized by an alter-ation in the level of attention and awareness, whichdevelops over a relative short period of time, and rep-resents a change fromthe subjects baseline.2Clinically,delirium is an acute or subacute organic mental syn-drome characterized by a disturbance in attention (i.e.,reduced ability to direct, focus, sustain, and shiftAm J Geriatr Psychiatry 21:12, December 2013for prehospital measures of global cognition, physicalfunctioning, and medical comorbidity.21 The numberof days of delirium older patients experience duringan ICU admission is signicantly associated withmortality up to 1 year after admission after control-ling for severity of illness (p 0.001).22Delirious patients, when compared with patients

suffering from the same medical problem who do notdevelopdelirium, experience prolongedhospital stays,on average 5e10 days longer.20,23e26 Among elderlymedically ill inpatients incident delirium was associ-ated with an excess stay after diagnosis of 7.78 days,even after controlling for covariates.27 Studies havedemonstrated that delirium in the emergency depart-ment was an independent predictor of prolongedhospital length of stay (i.e., twice as long) comparedwith nondelirious patients, even after adjusting forconfounding factors.28 Signicantlymorepatientswhodevelop delirium in the hospital ultimately requireinstitutional postacute care, such as placement in askilled nursing facility (e.g., 16% versus 3%), even aftercontrolling for illness severity, activities of daily livingstatus, prior cognitive impairment, and fever.20,26

A review of available empirical evidence suggestedthat a substantial proportion of patients who survivedelirium were much more likely to experience long-term cognitive impairment4,29e40 and to requireinstitutional postacute care, such as placement in askilled nursing facility (e.g., 16% versus 3%), evenafter controlling for illness severity, activities ofdaily living status, prior cognitive impairment, andfever.26,41 The increased morbidity and extendedhospital care associated with delirium have beenassociated with greater care costs.24,42,43 The totalannual direct healthcare costs attributable to deliriumin the United States has been estimated to be as highas $152 billion.44

NEUROPATHOGENESIS OF DELIRIUM

Despite its high prevalence and high morbidity,much is not understood about delirium. Known risk1% in a study of 229 medically ill, hospitalizedelderly patients.20 Even years after it occurred,delirium continues to have a long-lasting impact,with a 3-year mortality of 75% for patients who had1191

ics olate tr Ac.g., soriest is ustatiomple

Neuropathogenesis of DeliriumFIGURE 1. Theories on the development of delirium. Schematpathophysiology of delirium and how they may remechanism or pathologic process (e.g., DA excess o(e.g., sleep deprivation, aging), or empirical data (edelirium, intraoperative hypoxia). Most of these theintersection and reciprocal inuence. In the end, iexplaining the etiology or phenomenologic manifebiochemical derangement and, ultimately, to the cofactors for delirium include advanced age, preexistingcognitive impairment (including dementia), medica-tions (especially those with high anticholinergicpotential), sleep deprivation, hypoxia and anoxia,metabolic abnormalities, and a history of alcohol ordrug abuse.4 For years it has been predicted thatmetabolic processes must be involved in its etiology,as described by Engel and Romano over 50 years ago:We thus arrive at the proposition that a derangementin functional metabolism underlies all instances ofdelirium and that this is reected at the clinical levelby the characteristic disturbance in cognitive func-tions.45 (p. 532) Therefore, we begin with the premisethat delirium is a neurobehavioral syndrome causedby the dysregulation of neuronal activity secondary tosystemic disturbances.1,45e47

Over the years, a number of theories have beenproposed in an attempt to explain the processes

1192f the interrelationship of current theories on theo each other. Each proposed theory has focused on a specich deciency theories), observational and experiential evidencepecic pharmacologic agents association with postoperativeare complementary rather than competing, with many areas ofnlikely that any one of these theories is fully capable ofns of delirium but rather that their interaction lead to thex cognitive and behavioral changes characteristic of delirium.leading to the development of delirium.1 Each pro-posed theory has focused on a specic mechanism orpathologic process (e.g., dopamine [DA] excess oracetylcholine [Ach] deciency theories), observa-tional and experiential evidence (e.g., sleep depriva-tion, aging), or empirical data (e.g., specicpharmacologic agents association with postoperativedelirium, intraoperative hypoxia). To date, however,no single unitary pathophysiologic mechanism hasbeen identied.This article summarizes the seven most prominent

theories hypothesized to explain the phenomenonof delirium and highlights areas of commonalitywith the intent of increasing an appreciation of theunderlying pathophysiologic mechanisms, help focusfuture research, and assist in developing prophylacticand treatment strategies. Most of these theories arecomplementary rather than competing (Fig. 1). Thus,

Am J Geriatr Psychiatry 21:12, December 2013

may mediate the cognitive impairments observed after the acutepresentation of delirium has resolved.

parasympathetic nervous system (PNS). Cytokineexpression within the CNS is represented byasterisks within the brain. Dotted lines representnegative regulatory pathways, and solid linesrepresent positive regulatory pathways. CRH:corticotrophin-releasing hormone; AVP: argininevasopressin; TRH: thyrotropin-releasinghormone; GnRH: gonadotropin-releasinghormone; ACTH: adrenocorticotrophin hormone;TSH: thyroid-stimulating hormone; T4: thyroxine;T3: triio-dothyronine; LH: luteinizing hormone;FSH: follicle-stimulating hormone; LC: locuscoeruleus; Al, Cl, A2, C2: brainstem adrenergicnuclei. (From Marques-Deak et al.377).

Maldonado4. It is possible that instead of causing cognitive decits or dementia,delirium (and its underlying causes) only serve as a catabolic agentleading to an acceleration of normal physiologic cerebral agingmechanisms leading to dementia.

5. It is also possible that an episode of delirium simply unmaskssubtle cognitive decits already present but not yet identied.

Notes: Source: Maldonado222TABLE 1. Mechanisms Mediating Delirium and CognitiveImpairment

1. A number of factors and mechanisms leading to delirium may alsodirectly cause CNS damage and neuronal dysfunction and thusmediate both the manifestations of delirium and long-termcognitive impairment (e.g., cytokine release and otherneuroinammatory mediators; decrease perfusion andoxygenation leading to decreased cerebral oxidative metabolism;changes in BBB permeability; hypercatabolic states; water andelectrolyte imbalances; excessive GC levels and other HPA axisdysfunctions; melatonin and sleepewake cycle abnormalities).

2. Pharmacologic agents used either to treat the underlying causes ofthe delirium (e.g., steroids, calcineurin inhibitors, other immuno-suppressant agents, DA) or those agents used to treat delirium(e.g., DA blocking agents, benzodiazepines) may themselves leadto neuronal damage in a fragile brain.

3. Any mechanism listed above may themselves lead to alterations inneurotransmitter concentration or receptor sensitivity, whichitself may underlie the different symptoms and clinical pre-sentations of delirium and/or long-term cognitive dysfunction.Thus, the same mechanisms that cause the substrate for deliriumit is likely that none of these theories by themselves isfully capable of explaining the etiology or phenom-enologic manifestations of delirium but rather thattwo or more of these, if not all, act together to lead tothe biochemical derangement we know as delirium(Table 1).

NEUROINFLAMMATORY HYPOTHESIS

The Neuroinammatory Hypothesis (NIH) pro-poses that acute peripheral inammatory stimulation(from infectious, surgical, or traumatic etiologies)induces activation of brain parenchymal cells andexpression of proinammatory cytokines and inam-matorymediators in the central nervous system (CNS)that induce neuronal and synaptic dysfunction andsubsequent neurobehavioral and cognitive symptomscharacteristic of delirium (Fig. 2).48e53 Others havepreviously demonstrated how the brain monitors thepresence of peripheral inammation and how uponexposure to infection or inammation sick individualsdevelop nonspecic physiologic (e.g., fever, pain,malaise, fatigue, and anorexia) and behavioral

Am J Geriatr Psychiatry 21:12, December 2013FIGURE 2. Schematic illustration of neural immuneconnections. Immune signaling of the CNS viasystemic routes and the vagus nerve (Vagus n.)and CNS regulation of immunity via the HPA,hypothalamicepituitaryethyroid, andhypothalamicepituitaryegonadal axes and thesympathetic nervous system (SNS) and(e.g., lethargy, depressed mood, social withdrawal,cognitive loss, and anhedonia) changes known assickness behavior.54e59

Thus, ultimately the NIH suggests that deliriummay represent the CNS manifestation of a systemicdisease state that has indeed crossed the bloodebrain

1193

ferenleus tthe rus ams nerpresoughinaion oion bandresp

Neuropathogenesis of DeliriumFIGURE 3. Functional anatomy of the inammatory reex. Afvagus nerve to the nucleus of the solitary tract (nucoutow centers of the autonomic nervous system,neurons and the vagal motor neurons in the nuclearrives at the coeliac ganglion from either the vagusympathetic trunk. Stimulating the vagus nerve supproinammatory cytokine release in the spleen thra7 (a7nAChR). Note that after the activation of theare also relayed to the nuclei controlling the functadrenal gland. This provides an important connectcompensatory signals to adjust immune responseschronically modulate innate and adaptive immunebarrier (BBB).1 Many of the circumstances associatedwith a high incidence of delirium (e.g., infections,postoperative states) may be also associated with BBBintegrity compromise. It is theorized that severalillness processes (i.e., trauma, infections) and surgicalprocedures may introduce triggering factors leadingto the activation of the inammatory cascade: use ofanesthetic agents, extensive tissue trauma, presence offoreign organisms or substances, elevated hormonelevels, blood loss and anemia, blood transfusions, useof extracorporeal circulation, hypoxia, ischemia andreperfusion, formation of heparineprotamine com-plexes, and microemboli formation and migration(reviewed by Maldonado1).Others have demonstrated two pathways through

which immune signals from the periphery are trans-duced to the brain (Fig. 3): the neural pathway andthe humoral pathway.60 In the neural pathway,

1194t (sensory) neural signals to the brain stem are relayed by theractus solitaries [NTS]). Polysynaptic relays then connect to theostral ventrolateral medullary (RVLM) sympathoexcitatorybiguus (NA), and the dorsal vagal motor nucleus. Outowve or the preganglionic efferent nerves, which originate in theses innate immune responses and down-regulatesa mechanism that depends on nicotinic Ach receptor subunit

mmatory reex by sensory input to the brainstem, the signalsf the HPA axis, which increases GC hormone release by theetween the neural networks that can acutely providethe humoral anti-inammatory mechanisms that can moreonses. (From Tracey378).peripherally produced pathogen-associated molecu-lar patterns and cytokines activate primary afferentnerves, such as the vagus nerve. The humoralpathway involves circulating pathogen-associatedmolecular patterns that reach the brain at the level ofthe choroid plexus and the circumventricular organswhere pathogen-associatedmolecular patterns inducethe production and release of proinammatory cyto-kines by macrophage-like cells expressing Toll-likereceptors. Several studies have demonstrated thatpatients develop delirium during acute medicalhospitalizations experienced elevation of C-reactiveprotein, interleukin (IL)-6, tumor necrosis factor-a,IL-1RA, IL-10, and IL-8 as compared with patientswho did not have delirium, even after adjusting forinfection, age, and cognitive impairment, suggestingan association between proinammatory cytokinesand the pathogenesis of delirium.61e63

Am J Geriatr Psychiatry 21:12, December 2013

MaldonadoA study of adult patients admitted to medicinewards showed that those who developed deliriumhad signicantly elevated levels of IL-6 (53% versus31%) and IL-8 (45% versus 22%) when comparedwith patients who did not develop delirium, evenafter adjusting for infection, age, and cognitiveimpairment.61 This is the rst study to show a rela-tionship between peripherally measured cytokinelevels and delirium as a symptom of sicknessbehavior in acutely admitted elderly. It also showedthat cognitive function can be impaired by a systemicinfection in patients with a neurodegenerativedisorder such as Alzheimer disease and that thiscognitive decline is preceded by raised serum levelsof IL-1h. Similarly, a recent study found that highpreoperative neopterin levels predicted delirium aftercardiac surgery in older adults, suggesting thatplasma neopterin levels may be a candidatebiomarker for delirium among this patient popula-tion.64 Neopterin is produced by human monocytes/macrophages upon stimulation with the cytokineinterferon-g and thus can serve as a maker for im-mune system activation.65

CNS resident cells react to the presence of periph-eral immune signals, leading to production ofcytokines and other mediators in the brain, cell pro-liferation, and activation of the hypothalamicsepituitaryeadrenal (HPA) axis (see NeuroendocrineHypothesis, below) through a complex system of in-teractions. These neuroinammatory changes causeBBB permeability disruption (as suspected by eleva-tions of S100 beta) and changes in synaptic trans-mission, neural excitability, and cerebral blood ow,leading to the neurobehavioral and cognitive symp-toms characteristic of delirium (e.g., disruption inbehavior and cognitive functions).48 S100 beta is acalcium-binding proteinwith cytokine-like properties,secreted primarily by astrocytes under metabolicstress, and is considered a putative biomarker of CNSdamage: Increased cerebrospinal uid (CSF) andserum S100 beta are linked with adverse CNS out-comes. Several studies have reported ndings ofelevated S100 beta levels among patients sufferingfrom either dementia66e68 or delirium.69e72 Further-more, aging and neurodegenerative disorders exag-gerate brain microglial responses after stimulationby systemic immune stimuli such as peripheralinammation and/or infection (see Neuronal AgingHypothesis, below).Am J Geriatr Psychiatry 21:12, December 2013During or after various disease processes ortrauma/surgery, leukocytes adhere to endothelialcells, which make up the bulk of the bloodebrainbarrier and become activated, leading to degranula-tion and the release of free oxygen radicals andenzymes, which in turn leads to endothelial cellmembrane destruction, disruption of cellecell adhe-sions, and increased endothelial permeability.73,74

This in turn causes extravascular uid shifts andthe development of perivascular edema in cerebraltissue, which leads to decreased perfusion and longerdiffusion distance for oxygen.48,75,76 These processesmay lead to such extensive perfusion impairmentthat the blood ow in individual capillaries becomesdisrupted; thus, systemic inammation as a responseto trauma or illness leads to microcirculatoryimpairment and subsequent ischemic injury (seeOxidative Stress Hypothesis, below). Among thepertinent neurotransmitters, Ach synthesis andrelease may be the most sensitive to this type ofhypoxic injury and other homeostatic changes in thebrain (see Neurotransmitter Hypothesis, below).77

Similarly, neuroinammatory injuries have alsobeen associated with imbalances in other neuro-transmitters including DA, serotonin (5HT), andnorepinephrine (NE).78

In response to traumatic and systemic events, thesystemic inammatory response is activated, causingmonocytes and macrophages to produce neopterin,cytokines, and reactive oxygen species, which can befound in the plasma, urine, and CSF of deliriouspatients.1,48,63,64,79e82 In addition, disruptions of theendothelial cells (as described above) may also leadto enhanced cytokine transport across the disruptedBBB and inltration of leukocytes and cytokines intothe CNS, producing ischemia and neuronal apoptosis(Fig. 4).48,83,84

Systemic inammation is common in liver failure,and its acquisition is a predictor of hepatic encepha-lopathy (HE) severity. Studies provide convincingevidence for a role of neuroinammation in liverfailure; this evidence includes activation of microglia,together with increased synthesis in situ of proin-ammatory cytokines (i.e., TNF, IL-1b, and IL-6). Theproposed liverebrain signaling mechanisms inliver failure include direct effects of systemic proin-ammatory molecules, recruitment of monocytesafter microglial activation, brain accumulation ofammonia, lactate and manganese, and altered1195

ium.matof recammeral

ent mCer

Neuropathogenesis of DeliriumFIGURE 4. Effect of Inammation on the development of delirthe CNS. The initial interaction of circulating inamneurovascular unit occurs through a vast number opermeability of the BBB. In addition to systemic inhypoxia, ischemia, and pain. Recognition of periphevents leading to microglia activation and subsequ(represented with dashed reciprocal arrows). (Frompermeability of the BBB.85,86 This provides anotherintersection of the NIH with the NeurotransmitterHypothesis (NTH) as the above changes maycontribute to the alteration in neurotransmitterfunctioning in cases of HE (e.g., increased DA, 5HT,and g-aminobutyric acid [GABA]).There is also mounting evidence that some proin-

ammatory cytokines that induce sickness behavioralso enhance activity of the ubiquitous indoleamine-2,3-dioxygenase (Fig. 5).87,88Activation of indoleamine-2,3-dioxygenase results in decreased tryptophan (TRP)levels, thus a reduction in 5HT and melatonin produc-tion, whereas there is a shift to the production ofkynurenine and other TRP-derived metabolites thathave neurotoxic effects.89e91

Systemic inammation, such as that caused byinjury (including surgery) or infection, has long beenrecognized as trigger for episodes of delirium,particularly in elderly or demented patients, eventhough their deliriogenic effect seems to be lessenedin younger and nondemented patients.51,62,74,92e99

1196Recognition and propagation of peripheral immune stimuli inry mediators (e.g., cytokines and lipopolysaccharide) with theeptors and is associated with an increased paracellularation, other factors affect the integrity of BBB includinginammatory stimuli in the BBB is followed by a cascade ofodulation of adjacent cells including astrocytes and neuronsejeira et al.48).Yet, even in younger and nondemented adultpatients, the severity of the patients injury orunderlying medical problem is signicantly directlycorrelated with the development of delirium.100,101

These facts suggest that either a low dose of precip-itant in a vulnerable patient or a high dose in thenonvulnerable may overwhelm the system and leadto delirium. Thus, it seems likely that more severesystemic inammatory responses are more likely toinduce delirium, but pre-existing pathology incognitive circuitry is a stronger predictor; thus, theinteraction between these two factors is key.102 TheNIH intersects with the Neuronal Aging Hypothesis(NAH) because it has been shown that microglia, themajor macrophage population of the brain, areprimed by prior neurodegenerative pathology torespond more robustly to systemic inammatorysignals.53,103 There is also an interaction with theHPA axis (see Neuroendocrine Hypothesis, below) ashormones help coordinate an animals physiologyand behavior to match its environment and maximize

Am J Geriatr Psychiatry 21:12, December 2013

renicess bs a rederiv

MaldonadoFIGURE 5. Metabolic Pattern of TRP, 5HT, melatonin, and kynuproinammatory cytokines cannot only induce sickn2,3-dioxygenase, leading to decient TRP levels, thuproductionof kynurenine andother neurotoxic TRP-survival; among the most important processes inu-enced by endocrine hormones are the immunologicand behavioral responses to infection.104,105

NEURONAL AGING HYPOTHESIS

The NAH suggests that the aging process andaccompanying physiologic changes constitute anindependent risk factor for delirium. The concept ofhomeostenosis implies that a functional elderlyperson may maintain health into old age but maybecome increasingly vulnerable to stress and illness

Am J Geriatr Psychiatry 21:12, December 2013and quinolinic acid. Evidence suggests that someehavior but also enhance activity of the ubiquitous indoleamine-duction in 5HT and melatonin production, and a shift to theedmetabolites. (Adapted fromStone et al.87&Darlington et al.88).because of a lack of physiologic reserve.106 Accord-ingly, aging is associated with age-related cerebralchanges in stress-regulating neurotransmitters,braineblood ow decline, decreased vascular den-sity, neuron loss (particularly in locus coeruleus andsubstantia nigra), and intracellular signal trans-duction systems.107e122 This likely explains why theaging process itself is associated with some degree ofcognitive decits and an increased risk of dementia(Table 1). The NAH may also explain why the elderlyseem to experience a greater chance of developingdelirium when challenged by physiologic distressthat is better tolerated by younger individuals.

1197

Neuropathogenesis of DeliriumFIGURE 6. Age and transition to delirium. An estimation ofthe probability of transitioning to delirium by ageindicates that the incremental risk is large forpatients aged 65 years and older. The probabilityof transitioning to delirium increased dramatically(by 2%) for each year of life after 65 years.Adjusted odds ratio: 1.01 [(1.00, 1.02)], p [ 0.03.(Pandharipande P et al.)101Although the estimated annual incidence of Alz-heimer disease in the population is 0.6% for personsaged 65 and older,123 among elderly patients in theICU, the probability of developing delirium increasesby 2% per year of age for each year after age 65(Fig. 6).101 Other studies conrm that older age is anindependent risk factor among medically ill andsurgical patients, with the increase in rate per yeardependent on the specic context in which incidenceis measured.124e128 The data also suggest a reciprocalrelationship between delirium and cognitive decline:Dementia is the strongest risk factor for deliriumamong older patients,27,92,129,130 and the developmentof delirium appears to increase the risk of cognitivedecline, including dementia.32

It is well established that individuals withcompromised cognitive ability preoperatively (e.g.,dementia) are at greater risk of delirium.131e134 Astudy of elderly subjects undergoing orthopedicsurgery demonstrated that the presence of dementiaincreased the incidence of postoperative deliriumfrom 32% to 100%.33 However, evidence suggeststhat decrements in higher-order cognitive functions,such as executive function (e.g., problem solving,

1198processing speed, planning, complex sequencing, andreasoning), may also predict postoperative deliriumin the absence of frank cognitive impairment.135,136 Astudy of nondemented elderly patients undergoingelective orthopedic surgery demonstrated that subtlepreoperative attention decits, as tested by digitvigilance and reaction time testing, were closelyassociated with postoperative delirium.137 In fact,these subtle changes predicted a four- to vefoldincreased risk of postoperative delirium for subjectsmore than one standard deviation above the samplemeans on these variables.Conversely, studies have demonstrated that among

elderly surgical patients, delirium is a strong inde-pendent predictor of cognitive impairment and theoccurrence of severe dependency in activities of dailyliving. In fact, 38 months after discharge from hospital,53.8% of surviving patients with postoperativedelirium continued to experience cognitive impair-ment, as compared with only 4.4% of nondelirioussubjects.35 Similarly, a prospective, matched, con-trolled cohort study of elderly hip surgery patientsdemonstrated that the risk of dementia or minimalmild cognitive impairment over a 30-month follow-upalmost doubled in inpatients with postoperativedelirium compared with those without delirium.36

The data suggest that many older hospital patientsdo not recover from delirium and that persistence ofdelirium is associated with adverse outcomes. Insome studies, the long-term outcomes (e.g., mortality,nursing home placement, cognition, function) ofpatients with persistent delirium were consistentlyworse than the outcomes of patients who hadrecovered from delirium.138 These ndings suggestthat delirium does not simply persist for a certaintime but also predicts a future cognitive decline withan increased risk of dementia. Table 2 provides a listof potential mechanisms associated with theincreased risk of delirium in the elderly.The observed increased levels of circulating

inammatory mediators (e.g., cytokines, acute phaseproteins) suggest that chronic neurodegeneration isaccompanied by an inammatory response charac-terized by chronic, but selective, activation of CNSmicroglial cells that are primed to produce exag-gerated inammatory responses to immunologicchallenges.48,53 Age-related changes in the immunesystem, known as immunosenescence, and increasedsecretion of cytokines by adipose tissue represent theAm J Geriatr Psychiatry 21:12, December 2013

MaldonadoTABLE 2. Potential Mechanisms Associated with the IncreasedRisk of Delirium in the Elderly

Neuronal loss particularly in locus coeruleus and substantianigra110,111

Changes in various neurotransmitter systems111 Age-related decline in white matter integrity, observed as increasesmajor causes of chronic inammation, a phenomenonknown as inammaging.139 This inammationmay contribute to disease progression through theproduction of inammatory mediators. The agingprocess is associated with a two- to fourfold increase

whereas ve suggested a positive relationship(Table 3). At least one study found a positive rela-

in water diffusion and volume of hyperintense white matterlesions, intergyral spans, and reduction in fractional anisotropy ofwater diffusion, correlated with a decline in the global and regionalcerebral glucose uptake109

Age-related decline in regional cerebral blood ow, particularly inthe anterior cingulate gyrus, bilateral basal ganglia, left prefrontal,left lateral frontal and left superior temporal and insular cortex, asmeasured by single-photon emission tomography119

Age-related changes in cerebral blood ow, likely associated withbrain microvascular pathologies:112e116,118

B Rarefaction of the microvasculature in some regions of the brainB Decreased vascular densityB Damaged microvessels, with associated microinfarcts and

microhemorrhages, likely due to peripheral arterial aging lead-ing to stiffening and dilation of the proximal aorta with trans-mission of ow pulsations downstream into the brain

B Decline in cerebrovascular angiogenesisB Impaired cerebral blood ow due to tortuous arterioles and

deposition of excessive collagen in veins and venules Age-related decline in cerebral metabolic rate of oxygen moremarkedly in bilateral putamen, left supratemporal, left infrafrontal,and left parietal cortices121

Decreased oxygen supply (e.g., hypoxia) leading to a decrease inredox activity, resulting in decreased ACh production158,194

Decreased cerebral oxidative metabolism158 Age-related changes in brain neurochemical activity:122B There is a signicant increase in soluble hexokinase activity with

age, due to an increased release of mitochondrially boundhexokinase.

B There is a negative correlation of the activity of fructose-6-hosphate kinase with age, particularly in brain cortex andputamen.

B There is a signicant decline of carbonic anhydrase (importantin the regulation of the PO2/PCO2 ratio in the brain tissue) withincreasing age. Thus, PCO2-dependent regulation of tissue pH,ionic transport processes, and cerebral blood ow regulationhave the tendency to become more and more unstable.

B There is a progressive, age-dependent decline in cAMP-dependent activity, most signicantly in brain cortex andthalamus, followed by hippocampus, amygdala, and globuspallidus.

Age-related cerebral changes in stress-regulating neurotransmitterand intracellular signal transduction systems53

Decreased ACh levels in plasma and CSF181,185,187,192,195,196B This is likely due to

- Decreased volume of Ach-producing cells associated withnormal aging (see Neurotransmitter Hypothesis)158

- Decreased ACh synthesis associated with aging158

- Increase in baseline levels of circulating inammatory medi-ators including cytokines and acute phase proteins (seeNeuroinammatory Hypothesis)48,52,60,61,108,191

Am J Geriatr Psychiatry 21:12, December 2013tionship between delirium with APOE genotype,interferon-g, and insulin-like growth factor-I and thatrecovery was signicantly associated with lack ofAPOE4 allele and higher initial interferon-g.145 Theonly published meta-analysis on the matter found apositive association between postoperative cognitivedysfunction and the APOEs4 allele.150

OXIDATIVE STRESS HYPOTHESIS

Hypoperfusion appears to induce chronic oxida-tive damage in tissues and cells, largely due tothe generation of reactive oxygen and reactivenitrogen species. Any condition that outpaces thecapacity of endogenous redox systems to neutralizesuch toxic intermediates could lead to a systemimbalance or to major compensatory adjustmentsthat rebalance the system. This new redox state isgenerally referred to as oxidative stress (Fig. 7).151

The Oxidative Stress Hypothesis (OSH), initiallyproposed by Engel and Romano,45 proposes that anumber of physiologic processes, such as tissuedamage, hypoxia, severe illness, and infections,may give rise to increased oxygen consumptionand/or decreased oxygen availability, with associ-ated increased energy expenditure and reducedcerebral oxidative metabolism, leading to cerebralin baseline levels of circulating inammatory medi-ators, including cytokines and acute phase pro-teins.52,60,61 Other factors may inuence the medicallyill elderly patient, such as lower cognitive reserves,lower metabolic capacity, increased sensitivity tomedications, and lower threshold to a medicationsanticholinergic effects. In fact, even though highlevels of postoperative pain and high opioid use wasfound to increase the risk for postoperative deliriumin all elderly patients, the highest incidence ofdelirium occurred among patients who had highpreoperative risk for delirium (e.g., lower baselinecognitive status, comorbid psychiatric symptoms).140

Finally, there has been some controversy regardingthe contribution of apolipoprotein E (APOE) to thedevelopment of postoperative cognitive dysfunction.Of the nine articles published to date on the topic,four found that APOE-E4 genotype was not associ-ated with postoperative cognitive dysfunction,141e144

145e1491199

TABLE 3. Association Between APOE genotype and POCD

Study Population Findings p

Abildstrom et al., 2004141

Prospective, N 967Patients aged 40 y, older undergoing

noncardiac surgery One week after surgery, the incidence ofPOCD was 11.7% in patients with the 4allele and 9.9% in patients without the4 allele.

Conclusion: unable to show a signicantassociation between APOE genotypeand POCD.

0.41

Adamis et al., 2007145

Prospective, N 16470 y/o hospitalized medically ill patients Recovery was signicantly (p < 0.05)

associated with lack of APOE4 allele andhigher initial IFN-g.

A model incorporating gender, APOE 4status, and insulin-like growth factor-Ilevels predicted recovery or not fromdelirium in 76.5% of cases, with asensitivity of 0.77 and specicity of 0.75.

It further found a positive relationshipbetween delirium with APOE genotype,IFN-g, and insulin-like growth factor-Ibut not with IL-6, IL-1, TNF-a, andleukemia inhibitory factor.

65 y scheduled to undergo

major non-cardiac surgery requiringanesthesia

The presence of one copy of the 4 allelewas associated with an increased risk ofearly postoperative delirium (28.3% vs.11.1%; p 0.005). Even after adjustingfor covariates, patients with one copy ofthe 4 allele were still more likely tohave an increased risk of earlypostoperative delirium (OR: 3.64; 95%CI: 1.51e8.77) compared with thosewithout the 4 allele.

0.005

Tagarakis et al., 2007143 Elderly adults undergoing cardiac bypasssurgery; excluded dementia

Study conrmed the high incidence ofcognitive decline and delirium aftercoronary surgery but did not supportthe role of the APOE 4 allele in theoccurrence of delirium.

NS

Van Munster et al., 2007144

N 415Acutely admitted patients aged 65 y to

the Department of Medicine The OR for carriers of an APOE 4 allelecompared with patients without anAPOE 4 allele for developing deliriumwas 1.17 (95% CI: 0.49e2.78) in thecognitively intact patients and 0.42 (95%CI: 0.14e1.30) in the cognitivelyimpaired patients. No relation existedbetween the total number of APOE 4alleles and the different deliriumsubtypes.

0.12

Ely et al., 2007146

N 53Mechanically ventilated intensive care unit

patients Using multivariable regression analysis toadjust for age, admission diagnosis ofsepsis or acute respiratory distresssyndrome or pneumonia, severity ofillness, and duration of coma, thepresence of APOE4 allele was thestrongest predictor of delirium duration(OR: 7.32; 95% CI: 1.82e29.51).

0.005

Zhang et al., 2008149

N 196Elderly patients (>60 y/o) scheduled for

major abdominal surgery requiringgeneral anesthesia

The presence of the 4 allele and low levelof education were both associated withan increased risk of EA (36.9% vs. 15.8%,p 0.005; 30% vs. 14.3%, p 0.01).After adjustment for covariates, thepatients with the copy of 4 allele wereshown to have a greater likeliness of an

0.01

(Continued)

1200 Am J Geriatr Psychiatry 21:12, December 2013

Neuropathogenesis of Delirium

pedic/nivers07

n aort

ron- g

Maldonadodysfunction and associated cognitive and behavi-oral symptoms of delirium. In other words[delirium is] the clinical expression of a cerebralmetabolic defect.45 (p. 531)

Some have found that oxidative stress and/orantioxidant deciencies increase damage to cerebraltissue and lead to cognitive decline with irrever-sible degeneration as sequelae of delirium.30,152e154

Among patients undergoing cardiopulmonary

TABLE 3. (Continued)

Study Population

Van Munster et al., 2009148

Meta-analysis, N 656Medical Department and OrthoTraumatology Department of U

Hospital from 2003 to 20

Bryson et al., 2011142

N 88Patients 60 y/o undergoing ope

repair

Source: Maldonado222

Notes: POCD: postoperative cognitive dysfunction; IFN-g: interfebypass surgery, researchers found alterations in thelevels of markers of oxidative stress. In this sample,patients who developed delirium demonstratedsignicantly lower preoperative levels of the antiox-idant enzyme catalase (normally produced by thebody as the primary endogenous defense againstfree radicaleinduced injury) compared with non-postoperative delirium patients.Also, clinical data exist that correlate poor

oxygenation and cerebral dysfunction. A study ofmedically ill ICU patients demonstrated that oxida-tive metabolic stress was present within 48 hoursbefore the onset of delirium, as evidenced by abnor-malities in three measures of oxygenation (i.e.,hemoglobin, hematocrit, and pulse oximetry) amongpatients who developed delirium.155 In the samesample, clinical factors associated with greateroxidative stress (e.g., sepsis, pneumonia) occurredmore frequently among delirious subjects. Similarly,studies have demonstrated a strong correlation

Am J Geriatr Psychiatry 21:12, December 2013between intraoperative O2 saturation and post-operative mental function.156,157 Even healthy controlsubjects may experience delirium after dropping theirPaO2 to 35 mm Hg.

158 In fact, intraoperative cerebraloxygen desaturation was found to be a signicantrisk factor for postoperative delirium among cardiacsurgery patients.159 In addition, operative cerebraloxygen saturation levels among patients undergoingabdominal surgery was an independent risk factor

160

Findings p

increased risk of EA (OR: 4.32; 95% CI:1.75e10.05).

ityThe OR for delirium adjusted for age,cognitive, and functional impairment ofs4 carriers compared with non-s4carriers was 1.7 (95% CI: 1.1e2.6). Fourstudies were added to the meta-analysis,which included 1,099 patients in total.The OR for delirium in the meta-analysiswas 1.6 (95% CI: 0.9e2.7) of s4 carrierscompared with non-s4 carriers. Thisstudy and meta-analysis suggest anassociation between delirium and theAPOE s4 allele.

0.04

ic Delirium predicted POCD at discharge(OR: 2.86; 95% CI: 0.99e8.27) but not at3 months. APOE 4 genotype was notassociated with either delirium or POCDafter adjustment for covariates.

0.625

; NS: nor signicant; OR: odds ratio; CI: condence interval.for postoperative delirium. Similarly, among agroup of patients undergoing cardiac surgery, thosewho developed postoperative delirium had lowerpre- and intraoperative cerebral oxygen saturationlevels, were older, and had lower preoperativehemoglobin levels compared with nonpostoperativedelirium patients.161

Animal studies suggest that neuronal susceptibilityto ischemic injury is not uniform. The basal ganglia,thalamus, Purkinje, layer 3 of the cortex, and thepyramidal neurons of the hippocampus are particu-larly vulnerable to hypoxia, but the degree of damagemay vary depending on the cause.162 A super-imposed global mild ischemic injury is often presentin critically ill patients, galvanizing the oxidativefailure. Indeed, patients in the critical care settingare particularly at risk to suffer the effects of hypo-perfusion resulting from a number of potentiallycontrollable extrinsic factors (e.g., intraoperativehypotension).

1201

chemzationsiblesducntracher nr inactivad freue to

Neuropathogenesis of DeliriumFIGURE 7. Mechanisms of brain injury after global cerebral isextracellular compartment due to cellular depolariin intracellular Ca2D. The cascade of events responinduction, whereby extracellular GLU efux is tranintracellular Ca2D overload, which leads to lethal iwith an increase in intensity and involvement of otcascades. Excess release of Ca2D and its intracelluladeleterious intracellular processes that result fromto cell membrane breakdown, arachidonic acid, anfragmentation of genomic DNA and energy failure dInadequate oxidative metabolism may be one ofthe underlying causes of the basic metabolic prob-lems initiating the cascade that leads to the devel-opment of delirium, namely the inability to maintainionic gradients causing cortical spreading depression(i.e., spreading of a self-propagating wave of cellu-lar depolarization in the cerebral cortex);163e168

abnormal neurotransmitter synthesis, metabolism,and release;169e177 and a failure to effectively elimi-nate neurotoxic byproducts.170,171,175

The OSH intersects with the NTH becausedecreased oxygenation causes a failure in oxidativemetabolism, which may be one cause of the problemsobserved in delirium, namely decreased oxygenation,which causes a failure in oxidative metabolism andleads to a failure of the ATPase pump system.179

When the pump fails, the ionic gradients cannot bemaintained, leading to signicant inuxes of sodium

1202ia. During cerebral ischemia, excess GLU exits into the, coupled with its impaired uptake, which results in increasesfor GLU excitotoxicity includes three distinct processes: (1)ed by receptors on the neuronal membrane to causeellular derangements; (2) amplication of the derangement,eurons; and (3) expression of cell death triggered by cytotoxicux is thought to be the primary trigger for a variety of complex,tion of catabolic enzymes such as phospholipases (which leade radical formation) and endonucleases (which lead tomitochondrial dysfunction). (From Harukuni & Bhardwaj379).(Na) followed by calcium (Ca2), whereas potas-sium (K) moves out of the cell.178,179 Some havetheorized that it is the excess inward ux of Ca2 thatprecipitates the most signicant neurobehavioraldisturbances observed in delirious patients.232,233 Theinux of Ca2 during hypoxic conditions is associ-ated with the dramatic release of several neuro-transmitters, particularly glutamate (GLU) andDA.178,179 GLU further potentiates its own release asGLU stimulates the inux of Ca2233,234 and accu-mulates in the extracellular space as its reuptake andmetabolism in glial cells is impeded by the ATPasepump failure.179 In addition, at least two factorsfacilitate dramatic increases in DA: First, the con-version of DA to NE, which is oxygen dependent, issignicantly decreased, and second, the catechol-o-methyl transferase enzymes, required for degrada-tion of DA, get inhibited by toxic metabolites under

Am J Geriatr Psychiatry 21:12, December 2013

transmission. The NTH stresses the fact thatthe cholinergic and dopaminergic systems interact not

Maldonadoonly with each other but with glutamatergic andGABA pathways.188 In fact, some pharmacologicagents (e.g., opioids) may cause delirium byincreasing DA andGLU activity while decreasing Achavailability.189 In general, the most commonlydescribed neurotransmitter changes associated withdelirium are reduced availability of Ach ( Ach);excess release of DA ( DA), NE ( NE), and/or GLU( GLU); and alterations (e.g., both a decreased andincreased activity depending on circumstances andetiological factors) in 5HT ( 5HT), histamine ( H1and H2), and/or GABA ( GABA) (Table 1), aspreviously reviewed by others (Table 4).1,4,108

The cholinergic system is one of the most importantmodulatory neurotransmitter systems in the brain,controlling activities that depend on selective atten-tion, which are an essential component of conscioushypoxic conditions, leading to even greater accumu-lation of DA.216 At the same time, 5HT levels fallmoderately in the cortex, increase in the striatum, andremain stable in the brainstem.235 Hypoxia also leadsto a reduced synthesis and release of ACh, especiallyin the basal forebrain cholinergic centers.236 Indeed,cholinergic neurotransmission is particularly sensi-tive to metabolic insults, such as diminished avail-ability of glucose and oxygen.237 ACh synthesisrequires acetyl coenzyme A, which is a key interme-diate linking the glycolytic pathway and the citricacid cycle. Thus, reduction in cerebral oxygen andglucose supply and deciencies in enzyme cofactorssuch as thiamine may induce delirium by impairingACh production.77,238 These changes have beenextensively reviewed and discussed elsewhere(Fig. 8).1

NEUROTRANSMITTER HYPOTHESIS

The NTH was proposed after clinical observationsthat delirium occurred after the use of substances(e.g., medications, toxins) that alter neurotransmitterfunction or availability; it was originally used toexplain delirium secondary to impaired cholinergicfunction.78,180e185 Some have postulated that deliriummay be seen as a temporary psychiatric disorderresulting from a reduced central cholinergic trans-mission, combined with an increased dopaminergic

108,186,187Am J Geriatr Psychiatry 21:12, December 2013awareness190 (the two key components in Diagnosticand Statistical Manual of Mental Disorders, Fifth Edition,criteria for diagnosing delirium). Adequate AChlevels are also essential for the regulation of rapid-eye-movement sleep, memory, and synchronization ofthe electroencephalogram, among others. One ofthe leading hypotheses is that delirium results froman impairment of central cholinergic trans-mission1,181,191e193 and is considered by some to be acommon denominator in delirium (or toxicemetabolic encephalopaties).194 Studies have demon-strated low levels of ACh in plasma and CSF amongdelirious patients.181,185,187,192,195,196 Many havedemonstrated a relationship between a drugs anti-cholinergic potential and its deliriogeniceffects,181,184,197e202 thus raising awareness of thepotentially signicant additive effect of medicationscommonly thought to have lowACh activity.185,203,204

Similarly, high levels for serum anticholinergicactivity198 have been associated with an increasedlikelihood of delirium among surgical205 and medi-cal206 inpatients. In fact, serum anticholinergic activ-ity levels may predict delirium,181 and resolvingdelirium has been correlated to decreasing serumanticholinergic activity levels.207 Of interest, somehave found detectable serum anticholinergic activitylevels in delirious patients not exposed to knownanticholinergic agents, suggesting that endogenousanticholinergic substances may be produced duringacute illness and could be implicated in the etiologyof delirium.195,208 In a group of hospitalized elderlypatients, the Anticholinergic Risk Scale was able topredict the all-cause mortality when factored in alongwith baseline cognitive impairment, in-hospitaldelirium, place of residence (e.g., home versusnursing home), and length of hospital stay.209

Animal studies have revealed impairment incholinergic neurotransmission in several models ofencephalopathy and delirium, including hypoxia,nitrite poisoning, thiamine deciency, hepatic failure,carbon monoxide poisoning, anesthesia, selectiveintracranial atropine injection, physical immobiliza-tion, and hypoglycemia.186,210,211 Finally, potentialanimal models mimicking the decreased mobility ofcritically ill patients have demonstrated that immo-bilization may cause widespread reduction in AChlevels.210,211 Here the NTH intersects with the NAHbecause aging itself is associated with age-relatedcerebral changes in stress-regulating hormones and1203

eliriuTH ove ane.g.,xygencerebr clinrotrae inquateKD), DA,

Neuropathogenesis of DeliriumFIGURE 8. Oxidative stress and neurotransmitter theories of dtheorized intersections between the OSH and the Noutcomes, which may explain the complex cognitiproposes that a number of physiologic processes (increased oxygen consumption and/or decreased oreduced cerebral oxidative metabolism, leading tosymptoms of delirium. The NTH was proposed aftesubstances (e.g., medications, toxins) that alter neuNTH because decreased oxygenation causes a failursystem, which leads to an inability to maintain adealterations (e.g., inux of NaD and Ca2D; efux ofavailability) of several neurotransmitters (e.g., GLUintracellular signal transduction systems. Aging isalso associated with a decreased volume of AChproducing cells and decreased cerebral oxidativemetabolism (see Oxidative Stress Hypothesis, above).Similarly, hypoxia is known to reduce the synthesisand release of ACh.212 In addition, delirium is asso-ciated not only with an unbalanced inammatoryresponse but also with a dysfunctional interactionbetween the cholinergic and immune systems (seeNeuroinammatory Hypothesis, above).213 In fact,acute systemic inammation is a major trigger for

1204m. A basic pathoetiologic model of delirium illustrates thef the delirium, demonstrating potential common biochemicald behavioral changes characteristic of delirium. The OSHhypoxia, severe illness, infectious processes) may give rise toavailability, with associated increased energy expenditure andral dysfunction and associated cognitive and behavioralical observations that delirium occurred after the use ofnsmitter function or availability. The OSH intersects with theoxidative metabolism, leading to a failure of the ATPase pumpionic gradients, which in turn leads to signicant electrolyte

and subsequent alterations (e.g., excess release or decreasedAch). (From Maldonado1).cholinergic hypoactivity and is thought to be impor-tant in cognitive dysfunction during delirium.214

Elevations of DA have long been suspected in thedevelopment of delirium.187,188,191 For example,studies have conrmed elevation of DAs metabolites(i.e., homovanillic acid) in the CSF of patients withfulminant HE.215 At least two factors facilitate dra-matic increases in DA. First, the conversion of DA toNE, which is oxygen dependent, is signicantlydecreased (allowing DA to accumulate). Second, thecatechol-o-methyl transferase enzymes, required for

Am J Geriatr Psychiatry 21:12, December 2013

TABLE 4. Theorized Neurochemical Mechanisms Associated with Conditions Leading to Delirium

Notes: likely to be increased or activated; likely to be decreased or slowed; no signicant changes; ( ) likely a contributor, exact mechanism is unclear; () likelynot to be a contributing factor; CVA cerebro-vascular accident; Sx surgery; ETOH alcohol; CNS-Dep central nervous system depressant agent; ACH acetylcholine; DA dopamine; GLU glutamate; GABA gamma-aminobutyric acid; 5HT 5-hydroxytryptamine or serotonin; NE norepinephrine; Trp tryptophan; Phe phenylalanine; His histamine; Cytok cytokines; HPA axis hypothalamic-pituitary-adrenocortical axis; NMDA N-methyl-D-aspartic acid; RBF regional blood ow; EEG electroencephalo-graph; Mel melatonin; Inam inammation; Cort Cortisol. Source: Adapted from Maldonado, J. R. (2008). Pathoetiological model of delirium: a comprehensive under-standing of the neurobiology of delirium and an evidence-based approach to prevention and treatment. Crit Care Clin 24(4): 789-856.

Am

JGeriatr

Psych

iatry21:12,

Decem

ber

20131205

Maldonado

Neuropathogenesis of Deliriumdegradation of DA, get inhibited by toxic metabolitesunder hypoxic conditions, leading to even moreamassment of DA.216

In delirious patients, several metabolic pathwayslead to signicant increases in DA under impairedoxidative conditions (reviewed elsewhere1). One suchpathway is when signicant amounts of DA arereleased and there is a failure of adequateDAreuptake.A second pathway is when the inux of Ca2 stimu-lates the activity of tyrosine hydroxylase,217 whichconverts tyrosine to 3,4-dihydroxyphenylalanine andleads to increased DA production and further un-couples oxidative phosphorylation in brain mitochon-dria.178 The outcome is a disruption of adenosinetriphospate production and the increased productionof toxic metabolites of DA (formed under hypoxicconditions) that inhibit the activity of the oxygen-dependent catechol-O-methyl transferase,212,216 themajor extracellular deactivator of DA, further leadingto high levels of DA. Finally, an increase in the ringrates of catecholamine neurons may further inducetyrosine hydroxylase synthesis, which leads to evenmore DA production.218

Elevation in DA availability may lead to some ofthe neurobehavioral alterations observed in deliriouspatients, via DAs direct effects and by potentiatingGLUs excitotoxic effects.1,187,216 DA may exert itsdeliriogenic effect by one of three mechanisms: (1) thedirect excitatory activity of DA (e.g., toxicity withsubstances known to increase DA release or avail-ability, such as amphetamines, cocaine, and DA); (2)DA enhances GLU-mediated injury;216 and (3) excessDA may itself induce apoptosis by mechanismsindependent of oxidative stress,219 which mayexplain why DA depletion by a-methyl-paratyrosinemay have a neuroprotective effect against hypoxicstress and injury220 and why DA blockade can beused to reducehypoxicdamage in thehippocampus.221

It is also important to note the growing body ofevidence demonstrating that antipsychotic agents areeffective not only as treatment of delirium,222e227

because a recent meta-analysis demonstrated thatantipsychotic agents are one of the few pharmaco-logic agents demonstrated to prevent delirium.228 Anew meta-analysis revealed no signicant differencesin efcacy among agents.229 Furthermore, the authorsconcluded that the available evidence does not indi-cate major differences in response rates betweenclinical subtypes of delirium, suggesting the potential1206efcacy of antipsychotic agents in the managementand prevention of hypoactive delirium as well as theagitated and mixed types. Thus, it is possible thatantipsychotic agents are not only effective in thesymptomatic management of the symptoms ofdelirium, but they also address the underlyingmassive DA surge associated with some forms ofdelirium, even the hypoactive type.4,222

GLU is the brains principal excitatory neuro-transmitter, yet excessive activation of N-methyl-D-aspartate (NMDA) receptors may lead to neuronaldegeneration and cell death.230,231 In at least onestudy of high-risk adults undergoing cardiac surgery,serum concentrations of NMDA, as measured byserum concentrations of NMDA receptor antibodies(NR2Ab), were predictive of severe neurologicadverse events (e.g., delirium, transient ischemicattack, or stroke) postoperatively.As in the case of DA, increased GLU availability

may be due to the massive Ca2 inux describedabove, associated with a number of physiologicconditions (e.g., hypoxia, hepatic failure), leading toGLU release. Excess GLU further stimulates Ca2

inux, which releases even more GLU.232e234 Nor-mally, GLU is released into the synapse and thenremoved by astrocytes and converted into glutamine,ending its action. Under oxidative conditions, how-ever, GLU accumulates in the extracellular space asits reuptake and metabolism in glial cells is impededby the ATPase pump failure.179 It appears that GLUrequires the presence of DA to exert some of its toxiceffects, namely Ca2-induced neuronal injury.171 Athigh levels, DA may cause enough depolarization ofneurons to activate the voltage-dependent NMDAreceptor, therefore facilitating GLUs neurotoxiceffects.239

It has been known for a while that high ammonialevels are a factor in the pathogenesis of delirium andHE in cirrhotic patients, but it may not be that wellknown that acute ammonia toxicity is mediated bythe activation of NMDA receptors.240 In fact, somehave suggested that drug-induced delirium wouldresult from such transient thalamic dysfunctioncaused by exposure to medications that interfere withcentral glutamatergic, GABAergic, dopaminergic,and cholinergic pathways at critical sites of action.188

Furthermore, GLU is metabolized by GLU-decarboxylase into GABA, which itself has beenimplicated in the development of the delirium.241Am J Geriatr Psychiatry 21:12, December 2013

MaldonadoGABA is the chief inhibitory neurotransmitter in thehuman CNS and plays a role in regulating neuronalexcitability throughout and the regulation of muscletone. Evidence suggests that GABA activity isincreased in some types of delirium but decreased inothers. For example, the increased GABAergic tonetheory of HE, a neuropsychiatric disorder associatedwith liver failure, proposes the role of increasedGABAergic neurotransmission in HE.242e245 Some ofthe supporting evidence comes from clinical experi-ence demonstrating that umazenil, a highly selectivebenzodiazepine antagonist at the GABA receptorcomplex, improved electrencephalographic activity,reversed coma, and improved symptoms of hypo-active delirium in cirrhotic patients and in some HEpatients.242,246 In addition, neurosteroids, synthesizedin the brain mainly by astrocytes independent ofperipheral steroidal sources (i.e., adrenals andgonads), are potent positive allosteric modulatorsof the GABAA receptor. As such, neurosteroidsstimulate inhibitory neurotransmission in the CNS byincreasing GABAergic tone, a suggested pathophysi-ologic mechanisms in HE.247

Conversely, decreased GABAergic activity has beendescribed in deliria caused by ethanol or CNS-depressant withdrawal248 and antibiotic-induceddelirium.249 There is also mounting evidence thatsome GABAergic substances (e.g., benzodiazepines)may themselves induce delirium due to a variety ofmechanisms, including101,188,250e252 (1) by interferingwith physiologic sleep patterns,253 (2) by interruptingcentral cholinergic muscarinic transmission at the levelof the basal forebrain and hippocampus,254e258 (3) byincreasing compensatory up-regulation of NMDA andkainite receptors and Ca2 channels,259 (4) by dis-rupting thalamic gating function,188 (5) by causingwithdrawal states upon their cessation, and (6) by dis-rupting the circadian rhythm of melatonin release.260

Acute NE release secondary to hypoxia or ischemialeads to further neuronal injury and the developmentor worsening of delirium.261 Specically, in cases ofalcohol withdrawal, excess noradrenergic activitydrives most of the symptoms (e.g., diaphoresis,tachycardia, increased blood pressure, restlessness,anxiety, agitation, tremors). Under normal circum-stances the a2-receptor inhibits the ring of presyn-aptic NE neurons. Yet, evidence suggests that duringalcohol withdrawal, signaling at the a2-receptor maybe less sensitive, resulting in an inability of theAm J Geriatr Psychiatry 21:12, December 2013noradrenergic system to regulate its ring.262,263 Inaddition, alcohol withdrawal causes an up-regulationof GLU transmission in the locus coeruleus (the majornoradrenergic nucleus of the brain), increasing theactivity of the noradrenergic system.264 This maycontribute to the autonomic instability, behavioralagitation, and psychosis seen during alcohol with-drawal delirium.265 Furthermore, catecholamines canenhance the activity of the bed nucleus of the striaterminalis neurons that may in turn increase theexcitability of glutamatergic bed nucleus of thestria terminalis neurons that project to the ventraltegmental area.266

Randomized clinical trials have demonstrated thatselective a2 agonist agents (e.g., dexmedetomidine)substantially decreased the incidence of postoperativedelirium compared with GABAergic agents.4,14,267,268

Similarly, a2 agonist agents have shown neuro-protective qualities by suppressing circulating cate-cholamine levels during cerebral ischemia.269

Alterations in 5-HT activity, both elevated anddecreased, have been linked to delirium developmentin various clinical populations. Normal 5HT synthe-sis and release in the brain is dependent on theavailability of its precursor TRP (Fig. 5). Reduced5HT levels have been identied in patients sufferingfrom hypoxia, infections and sepsis, alcohol with-drawal delirium, delirium associated with levodop-amine use for the treatment of Parkinson disease,immobility, catabolic states, and postoperativedelirium, among others.191,270 Others have found anassociation between low 5HT levels associated withhyperactive delirium.271 In fact, the sudden discon-tinuation of 5HT reuptake inhibitors has been asso-ciated with various neuropsychiatric syndromes,including delirium.272

Decreased TRP availability may lead to a reductionin 5HT.270,273,274 All large neutral amino acids (i.e.,phenylalanine [PHE], TRP, eucine, isoleucine, methi-onine, tyrosine, valine) compete to enter the brainthrough the same saturable carrier. Therefore, as theconcentration of one increases, CNS entry of otherLNAAs conversely decreases.275 PHE has the addi-tional interesting property of conversion to neuro-toxic metabolites and competes with TRP for entryinto the brain.276 Once it enters the brain, PHE com-petes with TRP and tyrosine for metabolism, viahydroxylation.270 Studies have demonstrated thatelevations of the PHE/LNAA ratio are independently1207

lopathy have also been found to experience increasedlevels of PHE and PHE metabolites in the plasma and

Neuropathogenesis of DeliriumCSF. Studies have found that an increased ratio offree-to-bound TRP enhances its availability to braintissue, which in turn increases 5HT synthesis, thusprecipitating HE.281

Conversely, elevated5HT levelshavebeendescribedamong patients suffering from 5HT syndrome, HE,and clozapine-induced delirium.191,215,282,283 At leasttwo reports suggest that selective 5HT3-type receptorantagonists may be effective in the treatment ofagitated postoperative delirium.284,285 Similarly,hepatic dysfunctionmay lead to decreasedmetabolismof precursor amino acids (i.e., PHE, tyrosine), whichmay lead to increases in TRP availability, whichleads to increases in 5HT. In fact, elevation in5-hydroxyindoleacetic acid levels has been associatedwith HE and in patients suffering from hypoactivedelirium.46 Figure 5 shows the relationship betweenTRP, 5HT, and kynurenic acid metabolism.Histamine receptors A1 and A2 are known to affect

the polarity of cortical and hippocampal neurons286

and that both increased and decreased histaminelevels may lead to delirium. It is well known thatpharmacologic antagonism of either receptor is suf-cient to cause delirium.287 Others have suggestedthat during surgical stress and hypoxia, there may bean excessive release of histamine, which may lead todelirium.288

A detailed review of the studies supporting thevarious neurotransmitter ndings have been pub-lished elsewhere and are not repeated here.1 TheNTH intersects with the OSH because the abnor-malities in neurotransmitter concentration or receptorsensitivity may underlie the different symptoms andclinical presentations of delirium that are caused by adecreased in cerebral oxidative metabolism.

NEUROENDOCRINE HYPOTHESIS

The Neuroendocrine Hypothesis suggests thatdelirium represents a reaction to acute stress, medi-ated by abnormally high glucocorticoid (GC) levels,associated with postoperative delirium.274,277,278

Studies of elderly medically ill patients suggest thatan elevated plasma PHE/LNAA ratio during acutefebrile illness is associated with delirium.191,277 Simi-larly, patients with hepatic279 and septic280 encepha-1208which induce a general vulnerability in brain neuronsby impairing the ability of neurons to survive aftervarious metabolic insults.289,290 GCs are steroid hor-mones that modulate metabolism, salt balance,development, reproductive processes, and immunefunction.104 Although acute elevations of GCsenhance some immune functions, such as leukocyteinltration at the sites of injury, chronic elevationsinduce leukocyte apoptosis, reduce proinammatorycytokine release, and generally suppress immuneactivity.Stress activates the HPA axis: Stressors (through

inputs from the brainstem nuclei and the amygdala)activate the paraventricular nucleus of the hypothal-amus resulting in the release of corticotrophin-releasing hormone, which (through the hypophysealportal system) acts on the pituitary gland, inducingthe release of adrenocorticotrophic hormone, whichpromotes GC (including cortisol) release from theadrenal cortex. Under normal circumstances, GCs actto aid the body in coping with the demands imposedby stress exposure, mobilizing energy stores, andsuppressing nonvital body functions (e.g., inamma-tory responses and reproduction) (Fig. 9).81,291e293

However, plenty of scientic evidence demon-strates that GCs, the adrenal steroids secreted duringstress, can have a broad range of deleterious effects inthe brain.292 Recent data demonstrated that repeatedor prolonged exposure to GCs has a negative impacton brain function and provide evidence suggestingsuch exposure may contribute to age-related cogni-tive decline.81,294 Growing evidence suggests thatGCs may have proinammatory effects in the brainand can enhance neuroinammation at multiplelevels in the pathway that link lipopolysaccharideexposure to inammation.295 In fact, GCs have beenshown to enhance ischemic and seizure-inducedneuronal injury.296 A number of mechanisms havebeen postulated to explain how excess GC release cancompromise the neurons ability to survive variousneurologic insults (e.g., seizures, ischemia), whichmay lead to or exacerbate cell death (Table 5).The above may explain how GCs contribute to the

pathogenesis of delirium, especially in theelderly.292,294,297 The hippocampus, derived frommedial regions of the telencephalon, is part of thelimbic systemandplays important roles in informationencoding, related to short-term and long-term mem-ory, and spatial navigation.298 The hippocampus alsoAm J Geriatr Psychiatry 21:12, December 2013

MaldonadoFIGURE 9. Neuroimmune circuits in delirium. The HPA axisand inammation. Various stressors can activatethe HPA axis. The hypothalamus is stimulated tosecrete corticotrophin-releasing hormone (CRH),which leads to adrenocorticotrophic hormone(ACTH) secretion into the peripheral circulation.ACTH in turn triggers adrenal GC release andproduction. The CRH system is inhibited by GCs ina negative feedback loop. Tumor necrosis factor(TNF)-a and IL-1 are produced from inammatorysites and are potent activators of the HPA axis. IL-6acts synergistically with GCs to stimulate thehepatic secretion of acute phase proteins.Although GCs are widely known for their anti-happens to contain the highest concentration of GCreceptors of any brain region and thus may be amajor target for the negative effects of excessive GClevels. Current evidence suggests that hippocam-pal malfunction occurs relatively early during themetabolic stress environment, leading to delir-ium,102,290,291,293,299e308 and that the highly catabolicGCs induce a general metabolic vulnerability in theseneurons and thus compromise their ability to survivevarious toxic insults.309 Hippocampal injury mayexplain some of the attentional decits and memory

inammatory actions, (L), more recentlyproinammatory effects have also repeatedlybeen reported, (D). (D): enhancing; (L):suppressing. (From Dinkel et al.291).

Am J Geriatr Psychiatry 21:12, December 2013dysfunction and errors in information processing(leading to confabulation) commonly seen in deliriouspatients.The hippocampal formation is of prime importance

for normal HPA axis shut-off, and a key abnormalityrelated to GCs excess in delirium seems to be anabnormal shut-off of the HPA axis as tested bythe dexamethasone suppression test.310 The loss ofnormal inhibition of adrenal steroidogenesis resultsin continuous secretion of peak amounts of cortico-steroids, which causes resistance to cortisol feedbackinhibition mediated by receptor loss in the hippo-campus.292,304,309 The increased GC availability

TABLE 5. Potential Mechanisms to Explain How Excess GCRelease Can Compromise Neurons Ability toSurvive Neurologic Insults that May Lead to orExacerbate Cell Death

Inhibiting glucose transport into neuron, thus inducing metabolicvulnerability292,301

Increase proinammatory cell migration, cytokine production, andeven transcription factor activity in the brain81

Amplifying the damaging cascade of GLU excess, calcium (Ca2)mobilization, and oxygen radical generation304,306,309,380,381

Inducing spine loss and dendritic atrophy, thus decreasingneuroplasticity382

Enhancing oxygen radical-mediated neurotoxicity383 Exacerbating the toxicity of other neurotoxins (e.g., adriamycin)whose mechanisms of action overlap GC pathways384

Impairing long-term potentiation303 Reducing hippocampal glial cell activation and proliferation385 Altering the expression and signaling of neurotropins, particularlybrain-derived neurotropic factor302,386

Exacerbating the breakdown of cytoskeletal proteins (i.e., tau)387 Impairing neurogenesis388associated with illness and trauma (e.g., burns, sur-gery) or exogenous steroid administration furtherdisrupt hippocampal function, which in turn mayfurther disinhibit the release of GCs, thus sustaininghigh levels of circulating cortisol.46 Studies havefound that patients experiencing postoperativedelirium had an impaired stress regulating systemwith signicantly elevated mean plasma cortisollevels compared with the preoperative baseline andnondelirious patients.311 Among patients withdelirium triggered by lower respiratory tractinfections, 78% were found by the dexamethasonesuppression test to be nonsuppressors comparedwith 14% of those with normal suppressionresponses.290 Demented patients with deliriumexhibited signicant differences in basal cortisollevels compared with demented, nondeliriouspatients.312 Furthermore, there was a strong linear

1209

to consistently precede onset of delirium in post-surgical cardiac patients323,324 and that ICU patientswith sleep deprivation were signicantly more likely

Neuropathogenesis of Deliriumto develop delirium than patients without sleepdeprivation.260,325

Melatonin plays important roles in multiple bodilyfunctions that may have potential implicationsregarding the development of delirium in the medi-cally ill: Melatonin has signicant chronobiotic effect(i.e., affecting aspects of biologic time structure), hasrelationship between delirium and dexamethasonesuppression test pathology irrespective of age andseverity of dementia: The greater the intensity ofdelirium, the greater the level of nonsuppression.312

Similarly, patients with poststroke delirium exhibi-ted signicantly greater activation of the HPA systemcompared with those without.310,313

DIURNAL DYSREGULATION ORMELATONIN DYSREGULATION

HYPOTHESIS

This hypothesis suggests that disruptions to the24-hour circadian cycle and the usual stages of sleepmay lead to disturbances in the integrity of sleep andthe physiologic sleep architecture.1,222 Sleep depri-vation has long been linked to the development ofdelirium314 and psychosis.315 Hospitalized patientsexperience severe alterations of the sleepewake cyclewith sleep loss, sleep fragmentation, and sleepewakecycle disorganization. The 24-hour internal clock(circadian pattern) is maintained by environmentalfactors, primarily light exposure, which affectsmelatonin secretion, and its disruption may lead tothe development of delirium.316

Others have demonstrated that sleep deprivationmay lead to signicant memory decits317 and tosymptoms of emotional imbalance, likely due todisconnection between the amygdala and the pre-frontal cortex.318 Similarly, chronic partial sleepdeprivation (i.e., sleeping limited to 4 hours pernight for 5 consecutive nights) may lead to cumula-tive impairment in attention, critical thinking, reac-tion time, and recall.319 Thus, cumulative sleep debtcan cause delirium in itself and can aggravate orperpetuate delirium and its associated cognitivedecits.320e322 Studies have found sleep deprivation1210sleepewake cycle regulatory effects, helps resetcircadian rhythm disturbances, is an effective freeradical scavenger with extensive antioxidantactivity (particularly nuclear andmitochondrial DNA)with strong antiapoptotic signaling function, hasextensive anti-inammatory activity, and possessessome antinociceptive and analgesic effects.326e328 Inaddition, melatonin reduces the afnity of GCsreceptors, prevents GCs inhibition of cell proliferation,and reduces the GC-induced neurotoxicity andapoptosis.329 These qualities may protect naturalmechanisms of learning andmemory.317,326e328,330e336

Melatonin also inhibits the aggregation of the amyloidbeta protein into neurotoxic microaggregates respon-sible for the neurobrillary tangles characteristic ofAlzheimer disease and prevents the hyper-phosphorylation of the tau protein.333,337e343 Datasuggest that melatonin may have potential implica-tions regarding the development of delirium in themedically ill and postoperative patient. Conversely,it has been theorized that the natural, age-relateddecline in brain melatonin may contribute to apro-amyloidogenic microenvironment in the agingbrain.337

Current evidence suggests that acute and chronicsleep deprivation is associated with decreasedproportions of natural killer cells,344 reducedlymphokine-activated killer activity,345 and reducedIL-2 production.345 Conversely, cytokines may play arole in normal sleep regulation by increasing nonerapid-eye-movement sleep and decreasing rapid-eye-movement sleep, and during inammatory events,an increase in cytokine levels may intensify theirimpact on sleep regulation.346 Moreover, sleepdeprivation may alter endocrine and metabolicfunctions, altering the normal pattern of cortisolrelease and contributing to alterations of GC feed-back regulation, glucose tolerance, and insulin resis-tance.347 We previously described the impact ofdisruption of GC regulation and delirium andcognitive function (see Neuroendocrine Hypothesis,above).Studies have demonstrated a relationship between

an irregular melatonin circadian rhythm (i.e., abnor-mally low serum levels of melatonin) and post-operative delirium.260,348 The administration ofmelatonin in ICU patients has been shown toimprove quality of sleep and prolongation of sleeptime,349 whereas others demonstrated that theAm J Geriatr Psychiatry 21:12, December 2013

presentations (e.g., hyperactive or hypoactive) andthat, depending on the degree of reversibility of those

Maldonadochanges, the delirium episode would have short orlong-term sequelae (or effects). Initially, it immedi-ately recognized two systems as potential culprits:the cholinergic and the GABAergic systems.The NDH highlights the role of the cortical cholin-

ergic system and associated projections in mediatingspecic attentional processing (i.e., sustained, selec-tive, and divided attention performance), arousal, andrapid-eye-movement sleepeassociated dreaming78,355

and that the selectivity of the behavioral effects ofcortical ACh is based on close temporal interactionswith converging sensory or associational corticalinputs.356 In fact, increased cholinergic activity inprefrontal regions is hypothesized to contribute tothe activation of the anterior attention system andassociated executive functions.355 Finally, the NDHprophylactic administration of low-dose exogenousmelatonin may decrease the incidence ofdelirium.350,351 Finally, data suggest that the use ofprophylactic melatonin decreases the incidence andseverity of sundowning and agitated behavior inelderly, demented individuals.332

Some have observed a relationship between themotoric delirium subtype and melatonin levels. Astudy of hospitalized elderly medical patients wereevaluated daily using the Dementia Rating Scale andurinary measures of 6-sulphatoxymelatonin, the chiefmetabolite of melatonin.352 A study found that dur-ing periods of hyperactive delirium, subjects haddecreased urinary 6-sulphatoxymelatonin levels,whereas patients with hypoactive delirium hadraised 6-sulphatoxymelatonin levels.353

NETWORK DISCONNECTIVITYHYPOTHESIS

The Network Disconnectivity Hypothesis (NDH)suggests that the heterogenicity of delirium pre-sentations is better explained by the action ofvarious factors (e.g., drugs or toxins) acting on spe-cic brain neurochemical systems.354 Thus, thishypothesis postulates that factors affecting differentneurotransmitter-specic projections, whether due toaging (e.g., degeneration), disease (e.g., inamma-tion), or pharmacologic agents (e.g., anticholinergicsubstances), will lead to different types of deliriumAm J Geriatr Psychiatry 21:12, December 2013suggests that the failure (e.g., hypo- or hyperactivity)of one system will undoubtedly affect others (i.e., theconnectivity principle).The NDH recognizes that the brain is a highly

organized and interconnected structure functioningto allow complex integration of sensory informationand motor responses and suggests that deliriumrepresents a variable failure in the integration andappropriate processing of sensory information andmotor responses. Thus, the NDH proposes thatdelirium results from an acute breakdown in networkconnectivity within the brain.357

Furthermore, the NDH suggests that two impor-tant factors determine a subjects vulnerability todelirium: (1) the baseline network connectivity(dened as the connectivity of neural networkswithin the brain before the precipitating insult pro-voking delirium), which is inuenced by mostrecognized nonmodiable delirium risk factors (e.g.,age, baseline level of cognitive functioning), and (2)the level of inhibitory tone, which will determinethe degree of change in network connectivity andis inuenced by modiable risk factors (e.g., meta-bolic abnormalities, sleep deprivation, infection andinammation, medication such as benzodiaze-pines).357 The model suggests that when these twofactors affect separate neuronal networks, to differentdegrees, they produce the various motoric pheno-types described (e.g., hyperactive, hypoactive,mixed). Thus, the form of delirium that ensues willdepend on how and which networks breaks down(dependent on both the individuals baseline networkconnectivity and the degree of change in inhibitorytone produced).Electroencephalographic and evoked potential

data may provide further support for this theory bysuggesting that the pathophysiology of at least someforms of delirium may have a subcortical compo-nent.358 More recently, a study using functionalmagnetic resonance imaging scans examined thecorrelations of blood oxygen levels between variousbrain regions in resting-state functional magneticresonance imaging scans during and after the reso-lution of delirium. Findings demonstrated a long-lasting disruption in reciprocity of the dorsolateralprefrontal cortex with the posterior cingulate cortexand a reversible reduction of functional connectivityof subcortical regions (e.g., thalamus) with the retic-ular activating system and with nuclei responsible for1211

through functional disconnection between differentanatomic structures.375,376 Finally, it intersects with

Neuropathogenesis of Deliriumforebrain Ach (i.e., the midbrain nucleus basalis) andDA (i.e., the midbrain ventral tegmental area) inner-vation.359 The persistence of these physiologic dis-ruptions, beyond the resolution of acute delirioussymptoms, may account for reported cognitiveproblems that often outlast the acute episode ofdelirium.The NDH intersects with the NTH because

GABAergic neurotransmission is implicated inincreasing the inhibitory tone, which may contributeto the development of delirium.357 Furthermore,GABAergic agents may further destabilize thesleepewake cycle, as they suppress orexinergicneural ring in the perifornical nucleus.360 Orexin(hypocretin) is a peptide produced in the lateralhypothalamus that strengthens the ascending retic-ular activating system, thus maintaining wakefulnessand preventing inappropriate transition intosleep.17,18,265,361 It also relates to the NIH becausesystemic inammation drives an up-regulation inexpression of GABAA receptors and an increasedGABA synthesis362,363 and suppresses orexinergicneuronal activity during the wakeful period.364

This model also intersects with the NAH because itsuggests that aging is associated with gray mattervolume, neurotransmission, and white matter integ-rity (which consists mostly of glial cells and myelin-ated axons that transmit signals from one region ofthe cerebrum to another and between the cerebrumand lower brain centers) and thus may representan anatomic correlate of functional connectiv-ity.357,365,366 A study assessing white matter integrityusing diffusion tensor imaging in patients whodeveloped delirium after cardiac surgery suggestedthat abnormalities in the deep white matter andthalamus could have accounted for the deliriouspatients vulnerability to postoperative delirium,compared with nondelirious subjects.367 A morerecent study, also conducted among elderly cardiacsurgery patients, demonstrated that the prevalence ofsevere cerebral white matter hyperintensities onmagnetic resonance imaging was signicantly higherin delirious patients and similarly concluded theselesions were likely one of the most important riskfactors for the development of delirium after cardiacsurgery.368

Aging is also associated with a reduction inGABAergic tone, thus allowing for increasedneuronal activity in certain brain regions, primarily1212the NIH because acute neuroinammatory reactionsaffect physiologic processes implicated in neuronaland synaptic function with consequent neurochem-ical disturbances and functional disconnectionbetween different anatomic structures.48

CONCLUSION

Delirium is a neurobehavioral syndrome caused bythe disruption of neuronal activity secondary tosystemic disturbances. To date, no single unitarypathophysiologic mechanism has been identied.Most existing theories on the etiology of delirium arecomplementary rather than competing. Thus, it islikely that none of these theories by themselvesexplains the full phenomenon of delirium but ratherthat two or more of these, if not all, act together tolead to the biochemical derangement we know asdelirium.A review of the available literature suggests a

number of factors that lead to a nal commonpathway associated with alterations in neurotrans-mitter synthesis and availability that mediates thebehavioral and cognitive changes observed indelirium. These alterations in neurotransmittersynthesis appear to provide a relatively satisfactorythe prefrontal cortex.369 With advanced age there is adown-regulation of several subunits of the GABAAreceptors (i.e., a1, a5, b3, g2); therefore, a stimulusthat increases the level of inhibitory tone may have arelatively greater effect and further breakdownnetwork connectivity.370 The aged brain also experi-ences reductions in orexin signaling,371e374 whichmay contribute to the uctuating arousal level seen indelirious states when exposed to various noxiousstimuli (e.g., infection, GABAergic agents, sleepdeprivation). Similarly, it intersects with the NTHbecause it specically implicates the GABA system(e.g., benzodiazepine or HE-induced delirium) andthe cholinergic system (e.g., anticholinergic delirium)as likely principal culprits.The NDH also intersects with the NTH because

data from anesthetic agent studies demonstrate thatsome core symptoms of delirium likely involvechanges in dynamic aspects of neuronal activityaffecting brains ability to integrate informationAm J Geriatr Psychiatry 21:12, December 2013

eren

19. Rothschild JL, Leape LL: The Nature and Extent of Medical

Maldonadoacute hospital population: predictors, prevalence and detection.BMJ Open 2013; 3(1). http://dx.doi.org/10.1136/bmjopen-2012-001772

7. Inouye SK: Delirium in hospitalized older patients. Clin GeriatrMed 1998; 14:745e764

8. Naughton BJ, Moran MB, Kadah H, et al: Delirium and othercognitive impairment in older adults in an emergency depart-ment. Ann Emerg Med 1995; 25:751e755

9. Cole MG: Delirium in elderly patients. Am J Geriatr Psychiatry2004; 12:7e21

10. Elie M, Rousseau F, Cole M, et al: Prevalence and detection ofdelirium in elderly emergency department patients. CMAJ 2000;163:977e981

11. Wiesel O, Klausner J, Soffer D, et al: [Post-operative delirium of theelderly patientan iceberg?]. Harefuah 2011; 150:260e263, 303

12. Vaurio LE, Sands LP, Wang Y, et al: Postoperative delirium: theimportance of pain and pain management. Anesth Analg 2006;102:1267e1273

13. Dyer CB, Ashton CM, Teasdale TA: Postoperative delirium. Areview of 80 primary data-collection studies. Arch Intern Med1995; 155:461e465Am J Geriatr Psychiatry 21:12, December 2013Injury in Older Patients: Executive Summary. Washington, DC,Public Policy Institute, AARP, 2000

20. Francis J, Martin D, Kapoor WN: A prospective study of deliriumin hospitalized elderly. JAMA 1990; 263:1097e1101

21. Curyto KJ, Johnson J, TenHave T, et al: Survival of hospitalizedelderly patients with delirium: a prospective study. Am J GeriatrPsychiatry 2