41.
DELIR IUM
N EU ROBIOL O G Y, CH A R AC T E R I S T IC S , A N D M A N AG E M E N T
José R. Maldonado
I N T RODUC T IO N
Epidemiologically, delirium is likely the most common psy-
chiatric syndrome found in the general hospital setting. Its
occurrence parallels the severity of the underlying medical
processes. Clinically, delirium presents as an acute or sub-
acute organic mental syndrome characterized by disturbance
of consciousness, global cognitive impairment, disorienta-
tion, perceptual disturbance, deficits in attention, changes
in psychomotor activity, disordered sleep–wake cycles, and
fluctuations in symptom severity, all representing a signifi-
cant change over the patient’s baseline level of cognitive and
behavioral functioning. Etiologically, delirium is a neurobe-
havioral syndrome caused by the transient disruption of
normal neuronal activity secondary to disturbances of sys-
temic physiology. The phenotypes of delirium (i.e., hyperac-
tive, hypoactive, mixed) are likely a function of which of the
various neurotransmitter systems predominates, which can
change due to progression of illness or the influence of envi-
ronmental and/or pharmacological agents.
Hippocrates provided a clear description of the syndrome,
but called it phrenitis, meaning an “acute inflammation of
mind and body” (Lipowski, 1991). In fact, phrenitis literally
means “inflammation of the mind or diaphragm” (Greek
phren = mind, diaphragm; itis = inflammation) (Sakai, 1991;
Frederiks, 2000). It was an ancient disease concept or descrip-
tive notion that denoted a disease consisting of a combination
of a mental disorder and fever (the term phrenes had a double
meaning, indicating both a mental and a somatic domain),
distinguishing phrenitis from conventional madness (Sakai,
1991; Frederiks, 2000).
The term delirium is reported to have been first used by the
lay Roman writer Celsus (a.d. 1) (Adamis et al., 2007) and
described in his compendium De Medicina as “a madness
which is both acute and happens in a fever; the Greeks call it
phrenitis” (Celsus, 1814, Chapter XVIII, p. 115). According
to Celsus, a fever becomes a phrenitis “when the delirium
begins to continue without interruption; or when the patient,
though he still have his reason, yet forms to himself some vain
images:  it is perfect, when the mind gives itself up to these
images” (Celsus, 1814). It was also Celsus who was the first
to declare that “not all deliria were reversible, but in some
cases, although the causes disappeared, patients continued
823
to be demented.” Delirium is derived from the Latin roots
de (meaning “away from”), lira (meaning “furrow in a field”),
and ium (Latin for singular), and literally means “a going
off the plowed track; a madness.” Celsus described the vari-
ous forms of delirium: “Now there are several kinds of it: for
among phrenitic people some are merry, others sad; other
grow outrageous, and do acts of violence. . . . Those, who are
more violent in their actions, it is proper to bind, lest they
should hurt either themselves or any other person.” Despite
Celsus’s introduction of the term delirium, the syndrome
continued to be referred to as “phrenitis.”
In the second century a.d., Aretaeus classified diseases
as “acute” or “chronic.” Among the known mental disorders
of the time, phrenitis and lethargy (lethargus) were the chief
acute diseases, and both were believed to be caused by fever
or poisons and thus reflecting brain disease (Lipowski, 1991).
The ancient description of phrenitis typically involved rest-
lessness, insomnia, and hallucinations, and appears consis-
tent with our current understanding of hyperactive delirium.
Conversely, lethargy, involving undue quietness and sleepi-
ness, is more akin to what we now call hypoactive delirium.
It was not until the sixteenth century that we find the
first description of delirium in the English medical litera-
ture, in the textbook The Method of Physick (Barrough, 1583).
Barrough called it “frenesie” and described it involved the
derangement of three main functions:  imagination, cogita-
tion, and memory; also featuring disturbed sleep.
Thomas Willis (1621–1675) described delirium as result-
ing from “fever, poisoning, hemorrhage, lack of sleep, and
drunkenness” (Willis, 1672). According to Willis, its core
features included “incongruous conceptions and confused
thoughts,” distorted visual perceptions, and disturbed behav-
ior. Willis referred to phrenitis as “an inflammation of the
whole sensitive soul and animal spirits” (Bynum, 2000).
It was not until the nineteenth century that the term phre-
nitis was replaced by the word “delirium,” in part due to the
efforts of Gerard van Swieten (1700–1772), Giovanni Battista
Morgagni (1682–1771), and Philippe Pinel (1745–1826)
(Byl  & Szafran, 1996; Frederiks, 2000). After the works of
these luminaries of medicine, the word phrenitis fell into
disuse, replaced by the term febrile delirium (Berrios, 1981).
In 1813, the British physician Thomas Sutton introduced
the term delirium tremens to designate delirium caused by
the withdrawal from “spirits, and in consequence of excess
or habitual indulgence in them” (Sutton, 1813). In modern
times, the term is exclusively applied to delirium resulting
from alcohol withdrawal (2001). In the 1817 German treatise
“The Dream and the Feverish Insanity” (Der Traum und des
fieberhafte Irreseyn), Greiner first pioneered the term “clouding
of consciousness” (Verdunkelung des Bewusstseins) to describe
the pathophysiological abnormality responsible for delirium,
and linking disordered consciousness to the phenomenon of
delirium (Greiner, 1817). This relationship was further devel-
oped by Jackson, who viewed delirium as a clinical manifesta-
tion of disordered consciousness (Jackson, 1932). To Jackson,
delirium represented a state of reduced consciousness that
ranged from the “slightest confusion of thought to coma,” and
was due to some degree of “dissolution” of the topmost layer
of the nervous centers and consequent release from inhibition
of the evolutionarily lower centers.
Some of the most significant contributions in the field of
delirium include the works of:
Engel and Romano (1959), who demonstrated that the
behavioral and cognitive manifestations of delirium were
a reflection of an underlying metabolic derangement,
and demonstrated the relationship between reduction of
brain metabolic rate and electroencephalogram (EEG)
slowing;
Blass et al. (1981), who first proposed that impaired cere-
bral oxidative metabolism resulted in reduced synthesis
of various neurotransmitters, particularly acetylcholine,
leading to some of the characteristic deficits; and
Lipowski (1967, 1980, 1987), considered by many the
contemporary father of the study of delirium, who, in his
early seminal work, summarized the history of the field,
highlighted the problem with nomenclature, character-
ized its clinical features, summarized known etiological
factors, suggested diagnostic criteria and management,
and influenced the course of delirium research for years.
Unfortunately, one of the factors contributing to confu-
sion about delirium is that there are many other terms used
by various medical disciplines to describe the same phenome-
non: for example, “ICU (intensive care unit) psychosis,” “sun-
downing,” “acute confusional state,” “toxic confusional state,”
“post-anesthetic excitement,” “acute postoperative psychosis,”
“post-cardiotomy delirium,” and “toxic-metabolic encepha-
lopathy.” The author personally prefers the term acute brain
failure, as this term is compatible with multiple etiologies and
settings, describes the acuteness of the onset, implies that the
most important organ in the body, the brain, is failing, and
this may be associated with significant cognitive decline, mor-
bidity, and mortality.
E PI DE M IOL O G Y OF   DE L I R I U M
Delirium’s prevalence surpasses that of all other psychiat-
ric syndromes in the general medical setting (Maldonado,
824 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
2008b; NICE, 2010). In fact, delirium has been reported to
be one of the six most common preventable conditions in
hospitalized elderly patients (Rothschild & Leape, 2000).
Some specific medical conditions and surgical procedures
are associated with a particularly high incidence of delir-
ium during the postoperative period and among critical
care unit patients (see Table 41.1). Terminally ill cancer
patients, patients with moderate to severe traumatic brain
injury, frail elders, and the critically ill are clinical popu-
lations generally recognized to have a high incidence of
delirium over their course of illness, due to factors such as
polypharmacy, comorbid medical conditions, and meta-
bolic dysfunction.
A recent study at a large general hospital found that
delirium had a point prevalence of one in every five inpa-
tients, with some variation in certain medico-surgical units
and populations (Ryan, et al. 2013). The frequency of delir-
ium varies from 15% to 60% in hospitalized general medi-
cal and surgical patients (Smith & Dimsdale, 1989; Francis,
Martin, & Kapoor, 1990; Levkoff et al., 1992; Schor et al.,
1992; Williams-Russo, Urquhart, Sharrock, & Carlson,
1992; Pompei et al. 1994; Parikh & Chung, 1995; van der
Mast & Roest, 1996; Elie, Cole, Primeau, & Bellavance,
1998; Bucht, Gustafson, & Sandburg, 1999; Fann Alfano,
Roth-Roemer, Katon, & Syrjala, 2007; Aldemir Ozen,
Kara, Sir, & Bac, 2001; Lepouse, Lautner, Liu, Gomis, &
Leon, 2006; Siddiqi, House, & Holmes, 2006; Hala, 2007;
Lundstrom et al., 2007; Maldonado, 2008b; Katznelson
et al., 2009; Maldonado et al., 2009; Tognoni et al., 2011;
Ryan et al., 2013). The rate of postoperative delirium has
been reported to range from 10% to 74% (Dyer, Ashton, &
Teasdale, 1995; Vaurio, Sands, Wang, Mullen, & Leung,
2006; Bruce, Ritchie, Blizard, Lai, & Raven, 2007;
Maldonado et al., 2009; Wiesel, Klausner, Soffer & Zold,
2011). Studies have demonstrated that up to 87% of criti-
cally ill patients develop delirium during their ICU stay (Ely,
Margolin, et al., 2001). Delirium is common among cancer
patients (Adams, 1988; Weinrich & Sarna, 1994; Morita,
Tei, Tsunoda, Inoue, & Chihara, 2001; Breitbart, Gibson, &
Tremblay, 2002; Centeno, Sanz, & Bruera, 2004; Gaudreau,
Gagnon, Harel, Tremblay, & Roy, 2005, Gaudreau, Gagnon,
Roy, Harel & Tremblay, 2007).
Among the elderly, studies have found that between
14% and 24% of elderly patients are admitted to the hos-
pital with delirium (Inouye, 1998). It is estimated that
new cases arise in 6%–46% of hospitalized elderly patients
(Inouye, 1993). In the emergency department, the preva-
lence in elderly patients is 9.6% (Elie et al., 2000). This data
is consistent with more recent studies demonstrating that
among those aged 85+ years, the prevalence of delirium is
about 10%, rising up to 22% in populations with higher
percentages of demented elders and as high as 70% for those
in long-term care facilities (de Lange, Verhaak, & van der
Meer, 2013).
The occurrence of delirium seems to be highest among
the very sick. Various studies have repeatedly demonstrated
that up to 87% of those critically ill develop delirium during
their stay in critical care units (Ely, Margolin, et al., 2001).
 Table 41.1 A COMPAR ISON OF THE INCIDENCE OF PSYCHIATR IC DISOR DER
IN THE GENER AL POPULATION AND DELIR IUM AMONG MEDICALLY ILL PATIENTS

MEDICAL POPULATIONS
UNDER STUDY PR EVALENCE OF DELIR IUM (%)

General Medicine Wards 9–24% (Ritchie et al., 1996; Valdes et al., 2000; Maldonado, Dhami, & Wise,
2003; Gonzalez, de Pablo, et al., 2004; Speed et al., 2007)

HIV/AIDS 30%–40% (Fernandez, Levy, & Mansell, 1989; Uldall et al., 2000)

Medical-ICU 60%–80% (Ely et al., 2001)

Post-Stroke 13%–48% (McManus et al., 2007)

Surgery

Postoperative Delirium 10%–74% (Vaurio et al., 2006; Wiesel, Klausner, et al., 2011)

General Surgical Wards 7%–52% (Dyer & Teasdale, 1995; Bucht, Gustafson, & Sandburgh, 1999)

Spine Surgery 12.5% (Kawaguchi et al., 2006)

Post-CABG 25%–32% (Nevin, Colchester, et al., 1989; Ebert & Herrmann 2001)

Post-Cardiotomy 50%–67% (Smith & Dimsdale, 1989; van der Mast & Roest, 1996; Maldonado
et al., 2009)

Abdominal Aneurysm Repair 33% (Benoit et al., 2005)

Geriatrics

Outpatient Minor (Cataract) 4.4% (Milstein et al., 2002)

Surgery

At Time of Hospitalization 10%–15%

In Nursing Homes 15%–60%

Frail Elderly Patient 60% (Francis et al., 1990)

Elective Hip or Knee Replacement 25% (Gustafson, Berggren, et al., 1988; Marcantonio, Flacker, et al., 2000)

Bilateral Knee Replacement 41% (Williams-Russo et al., 1992)

Femoral Neck Fracture Repair 65% (Gustafson, Bucht, et al., 1988; Marcantonio et al., 2000)

Oncology

General Prevalence 25%–40% (Weinrich & Sarna, 1994; Olofsson et al., 1996; Tuma & DeAngelis,
2000)

Palliative Care Units 26%–44% (Lawlor, Pereira, et al., 2000; Morita, Tei, et al., 2001; Lawlor, 2002;
Centeno, Sanz, & Bruera, 2004)

Bone Marrow Transplantation 73% (Fann et al., 2007)

Advanced Cancer Up to 85% (Lawlor et al., 2000)

Psychiatry
Among Psychiatric Patients 14.6% (Ritchie et al., 1996)

In addition, delirium is common among cancer patients
(Adams, 1988; Weinrich & Sarna, 1994; Morita et al., 2001;
Breitbart et al., 2002; Centeno, Sanz, & Bruera, 2004;
Gaudreau et al., 2005b). Among those with advanced can-
cer, delirium was diagnosed in 42% of patients on admission,
while it developed in 45% of hospitalized cancer patients
who were not delirious at the time of admission (Lawlor
et al., 2000). Similar to patients in critical care units, termi-
nal delirium occurs in up to 88% of patients dying of cancer
(Lawlor et al., 2000).
The large variation of incidence and prevalence data
reported by the various studies reflects differences in patient
populations studied, the severity of their medical conditions,
and the diagnostic criteria used.

41. D E L I R I U M • 825
 I M PAC T OF   DE L I R I U M
MOR BI DI T Y A N D MORTA L I T Y R E L AT E D
TO DE L I R I U M
Across many hospital settings (e.g., emergency department,
general medical ward, postoperative, critical care units) and
after controlling for multiple variables (e.g., demographics,
age, gender, illness severity, and comorbid medical condi-
tions), patients who develop delirium fare much worse than
comparable nondelirious cohorts (Vasilevskis, Han, Hughes,
Ely, 2012). A  recent meta-analysis, including 16 studies
(n  =  6410)  among critically ill patients, demonstrated that,
despite modern advances, when compared to their nonde-
lirious counterparts, hospitalized patients with delirium had
higher mortality rates (odds ratio [OR]: 3.22; 95% confidence
interval [CI]: 2.30–4.52); had longer lengths of stay in both
the ICU (weighted mean difference [WMD]: 7.32 days; 95%
CI: 4.63–10.01) and in the general hospital (WMD: 6.53 days;
95% CI: 3.03–10.03); spent more time on mechanical ventila-
tion (WMD:  7.22  days; 95% CI:  5.15–9.29); experienced a
higher rate (6X) of complications (e.g., acute respiratory extu-
bation, removal of catheter, cardiac arrhythmia; OR: 6.5; 95%
CI: 2.7–15.6); and were more likely to be discharged to skilled
placement (OR:  2.59; 95% CI:  1.59–4.21) (Zhang, Pan,  &
Ni, 2013).
Delirious patients, when compared with patients suf-
fering from the same medical problem who do not develop
delirium as a complication, experience prolonged hospital
stays—five to ten days longer on average (Francis, Martin, &
Kapoor, 1990; O’Keeffe & Lavan, 1997; Ely, Gautam, et al.
2001; Maldonado, Dhami, & Wise, 2003; Ely, Shintani et al.,
2004). Significantly more patients who develop delirium in
the hospital ultimately require institutional post-acute care,
such as placement in a skilled nursing facility (e.g., 16% for
delirium patients versus 3% for those without delirium),
even after controlling for illness severity, activities of daily
living (ADL) status, prior cognitive impairment, and fever
(Francis, Martin, & Kapoor, 1990; O’Keeffe & Lavan, 1997).
Psychiatric hospital inpatients with delirium had hospital
stays that were 62.1% longer than those of patients without
delirium (Ritchie, Steiner, & Abrahamowicz, 1996).
Among critically ill patients there is an independent
association between delirium present within 24 hours
after admission to the critical care unit with an increased
in-hospital mortality (i.e., 16.2% for delirious patients vs.
5.7% for nondelirious; van den Boogaard et al., 2010). Others
have reported that patients who developed delirium in criti-
cal care units experience higher mortality, both at 90  days
(11% vs. 3%; Pompei et al., 1994) and at 6 months (34% vs.
15%; Ely et  al., 2004; (See http://jama.jamanetwork.com/
article.aspx?articleid=198503)). A  recent meta-analysis of
clinical observational studies in critically ill patients revealed
that, when compared with their nondelirious counterparts,
delirious patients had higher mortality rate (OR:  3.22;
95% CI:  2.30–4.52); a higher rate of complications includ-
ing removal of catheter, self-extubation, reintubation, acute
respiratory distress syndrome, nosocomial pneumonia,
826 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
cardiopulmonary edema, and cardiac arrhythmia (OR:  6.5;
95% CI:  2.7–15.6); longer length of stay in both the criti-
cal care unit (weighted mean difference [WMD]: 7.32 days;
95% CI:  4.63–10.01) and hospital (WMD:  6.53  days; 95%
CI:  3.03–10.03); spent more time on mechanical ventila-
tion (WMD: 7.22 days; 95% CI: 5.15–9.29); and were more
likely to be discharged to skilled placement (OR: 2.59; 95%
CI: 1.59–4.21) (Zhang et al., 2013).
Delirium’s impact among the elderly is even greater. The
mortality rate for elderly patients in acute care hospitals is
much higher among those with delirium than those with-
out delirium: 8% versus 1% in one study (Francis, Martin, &
Kapoor, 1990). Among hospitalized, elderly, medically ill
patients, incident delirium was associated with an excess
stay after diagnosis of 7.78  days, even after controlling for
covariates (McCusker, Cole, et  al., 2003). Similarly, studies
have demonstrated that delirium in the emergency depart-
ment was an independent predictor of prolonged hospi-
tal length of stay (twice as long), compared to nondelirious
patients, even after adjusting for confounding factors (Han
et  al., 2011). Furthermore, the number of days of delirium
older patients experience while in the critical care unit is
significantly associated with time to death within 1  year
post-ICU admission, after controlling for age, severity of ill-
ness, comorbid conditions, psychoactive medication use, and
baseline cognitive and functional status (hazard p = 0.001;
hazard ratio, 1.10; 95% CI, 1.02–1.18); see Thoracic Society
Journals, http://www.atsjournals.org/doi/abs/10.1164/rccm.
2 0 0 9 0 4 - 0 537O C? u rl _ver =Z 39.8 8 -2 0 03 & r f r_ id=
ori:rid:crossref.org&rfr_dat=cr_ pub%3dpubmed&#.
VB9cea10wfg – Figure 2 Pisani, Kong, et al., 2009.
These bleak prognoses apply not only to severe cases; even
patients with prevalent subsyndromal delirium (SSD) have
been shown to experience longer acute-care hospital stay,
increased post-discharge mortality, more symptoms of delir-
ium, and a lower cognitive and functional level at follow-up
than patients with no SSD (Cole, McCusker, Dendukuri, &
Han, 2003). Among patients admitted to an intensive care
unit, when comparing no delirium versus SSD, the mortality
rate goes up from 2.4% to 10.6% (p < 0.001), ICU length of stay
goes up from 2.5 to 5.2 days (p < 0.001), and the overall hos-
pital length of stay goes up from 12.9 to 16.7 days (p < 0.001)
(Ouimet, Kavanaugh, Gottfried, & Skrobik, 2007).
C O G N I T I V E A N D BE H AV IOR A L S E QU E L A E
There appears to be a reciprocal relationship between delir-
ium and cognitive decline, with evidence demonstrating that
the presence of baseline cognitive deficits, including demen-
tia, lowers the threshold to develop delirium (Inouye, 1998;
Wahlund & Bjorlin, 1999; Litaker, Locala, Franco, Bronson, &
Tannous, 2001; McNicoll et  al., 2003; Benoit et  al., 2005;
Kalisvaart et al., 2006; Wacker, Nunes, Cabrita, & Forlenza,
2006; Smith, Attix, Weldon, Greene, & Monk, 2009; Franco,
Valencia, et al., 2010; Tognoni et al., 2011). Data suggests that
baseline dementia is the strongest risk factor for delirium
among older patients (Elie et al., 1998; McCusker et al., 2003;
McAvay et al., 2006; Inouye et al., 2007). On the other hand,
the development of delirium appears to increase the risk of
cognitive decline, including long-term cognitive impairment
(LTCI) and dementia (Rockwood et al., 1999; Morandi et al.,
2012). A number of pathways and mechanisms have been pro-
posed, yet, despite of some suggestive evidence, none has been
conclusively proven as causative (see Box 41.1).
Both clinical experience and research data suggests that
many patients do not fully recover (cognitively) from delir-
ium, particularly the elderly. The proportion of patients with
residual cognitive impairment reported in various studies
depends on the clinical population under study, the cause
of the delirium (when it is known), and the threshold and
criteria used for diagnosing. In fact, a substantial number
of patients who survive delirium are left with post-delirium
cognitive impairment (Wacker et al., 2006; Griffiths & Jones,
2007; Bickel, Gradinger, Kochs, & Forstl, 2008; Kat et al.,
2008; Maldonado, 2008b; Fong et al., 2009; MacLullich,
Beaglehole, Hall, & Meagher, 2009; Girard et al., 2010). Data
demonstrates that, when compared to controls, patients diag-
nosed with delirium are much more likely to experience LTCI
(Macdonald, 1999; Rockwood et al., 1999; Jackson, Gordon,
Hart, Hopkins, & Ely, 2004; Gunther, Jackson, & Ely, 2007).
A study of mechanically ventilated patients who experienced
in-hospital delirium found that at 3-month and 12-month
follow-ups, 79% and 71% of survivors had cognitive impair-
ment, respectively, with 62% and 36% being severely impaired
(Girard et al., 2010). These studies demonstrate that delirium
duration was independently associated with long-term cogni-
tive outcome (i.e., the longer the duration of delirium, the more
significant the cognitive deficits). For example, an increase
from 1 day of delirium to 5 days was independently associated
with nearly a 5-point decline (i.e., a one-half standard devia-
tion [SD] decline) in the cognitive battery mean score (95%
CI, –9.2 to –0.1).In fact, more than 20 prospective studies
(> 5,000 patients) during the last three decades demonstrate
a significant association between delirium and long-term
cognitive dysfunction (MacLullich et al., 2009; Witlox et al.,
2010). The studies suggest that about 40% of patients with
delirium develop some form of cognitive impairment when
followed up from about 3 months to 5 years after an episode
of delirium (Levkoff et al. 1992, McCusker, Cole, Dendukuri,
Belzile, & Primeau, 2001; Jackson et al., 2004; Witlox et al.,
2010). A comprehensive meta-analysis concluded that there is
greater cognitive impairment among patients with delirium
than matched controls; there is a higher incidence of dementia
in patients with a history of delirium; and that one of three
survivors of critical illness with delirium developed cogni-
tive impairment (Jackson et al., 2004). A systematic search

Box 41.1 MECHANISMS MEDIATING DELIRIUM AND COGNITIVE IMPAIR MENT

1. A number of factors and mechanisms leading to delirium may also directly cause CNS damage and neuronal dysfunction, and thus
mediate both the manifestations of delirium and long-term cognitive impairment, including:
a) cytokine release and other neuroinflammatory mediators;
b) decrease perfusion and oxygenation, leading to decreased cerebral oxidative metabolism;
c) changes in blood–brain barrier permeability;
d) hyper-catabolic states;
e) water and electrolyte imbalances;
f) excessive glucocorticoid levels and other HPA-axis dysfunctions;
g) melatonin and sleep-wake cycle abnormalities
h) genetic factors (e.g., some have described an association between APOE4 and prolonged delirium duration)
i) Amyloid-β deposition (e.g., increased amyloid-β levels have been associated with subjective cognitive dysfunction in critical illness
survivors)
2. Pharmacological agents used to treat the underlying causes of the delirium (e.g., steroids, calcineurin inhibitors, other immunosup-
pressants, dopamine) or the agents used to treat delirium (e.g., dopamine-blocking agents, benzodiazepines) may themselves lead to
neuronal damage in a fragile brain.
3. Any of the mechanisms listed above may themselves lead to alterations in neurotransmitter concentration or receptor sensitivity,
which itself may underlie the different symptoms and clinical presentations of delirium and/or long-term cognitive dysfunction.
Thus, the same mechanisms that cause the substrate for delirium may mediate the cognitive impairments observed after the acute
presentation of delirium has resolved.
4. It is possible that instead of causing cognitive deficits or dementia, delirium (and its underlying causes) only serves as a catabolic
agent leading to an acceleration of normal, physiological, cerebral aging mechanisms leading to dementia.
5. It is also possible that an episode of delirium simply unmasks subtle cognitive deficits already present, although not yet identified.

41. D E L I R I U M • 827
of studies published between January 1981 and April 2010
found that delirium is associated with an increased risk of
death compared with controls (38.0% vs. 27.5%); that patients
who experienced delirium were at increased risk of institu-
tionalization (33.4% vs. 10.7%); and that patients who expe-
rienced delirium were at increased risk of dementia (62.5% vs.
8.1%) (Witlox et al., 2010).
Some have postulated four possible putative mechanisms
or explanations for the relationship between delirium and
long-term cognitive impairment: (1) that delirium is a marker
of chronic progressive pathology but unrelated to any progres-
sion; (2) that delirium is a consequence of acute brain dam-
age, which is also responsible for a “single hit” or triggering of
active processes causing LTCI; (3) that delirium itself is the
cause of the LTCI; and/or (4)  that pharmacological agents
used for the treatment of delirium or other conditions are
the cause of the LTCI associated with delirium (MacLullich
et al., 2009).
There appears to be a “dose effect” of delirium on cognition,
with a longer duration of delirium found to be associated with
worse average performance on neuropsychological testing at 3
and 12 months follow-up (p = 0.02 and p = 0.03, respectively)
even after adjusting for age, education, preexisting cognitive
function, severity of illness, severe sepsis, and exposure to sed-
ative medications in the critical care unit (Girard et al., 2010).
An increase from 1 to 5 days of delirium was independently
associated with a 7-point decline in the cognitive battery
mean score at 12 months follow-up (p = 0.03) (Girard et al.,
2010). Various factors may be associated with this dose effect:
neuroinflammatory changes; changes in brain–blood barrier
permeability; the effect of psychoactive medications; anemia,
hypoxia, anoxia, and other low-perfusion states; or systemic
organ failure (Maldonado, 2008a, 2013).
Yet, whatever the cause, there appears to be an association
between longer duration of delirium and greater brain atro-
phy as measured by a larger ventricle-to-brain ratio at hospi-
tal discharge (p = 0.03) and at 3-month follow-up (p = 0.05)
(Gunther et  al., 2012). As expected, greater brain atrophy
(higher ventricle-to-brain ratio) at 3  months was associated
with worse cognitive performance (executive functioning
and visual attention) at 12 months (p = 0.04) (Gunther et al.,
2012). Others have found that the longer the duration of
delirium, the greater the degree of white matter disruption
in the genu of the corpus callosum and anterior limb of the
internal capsule (p = .01), up to 3 months after hospital dis-
charge (Morandi, Rogers, et al., 2012).
A prospective study followed older hospitalized delirious
patients who suffered from prevalent or incident delirium
detected during the first week of hospitalization for up to a
year and found that 39%, 38.5%, and 48.9% of patients with
dementia met delirium criteria at discharge, 6-month, and
12-month follow-ups, respectively, compared to 11.1%, 8.8%,
and 14.8% of patients without dementia (McCusker, Cole,
et al., 2003). However, cognitive improvement during hospi-
talization and long-term changes in the number of symptoms
were remarkably similar in the two groups, with inattention,
disorientation, and impaired memory being the most com-
mon symptoms.
828 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
Another study found that only 4% of delirious patients
experienced full resolution of all symptoms of delirium
before discharge from the hospital (Levkoff et  al., 1992).
In this case, after following this sample longitudinally, an
additional 20.8% experienced resolution of symptoms by
the third month after hospital discharge, with an additional
17.7% by the sixth month after discharge from the hospital.
Among elderly hip surgery patients, delirium was a strong
independent predictor of cognitive impairment and the
occurrence of severe dependency in ADL. In fact, 38 months
after discharge from the hospital, 53.8% of the surviving
patients with postoperative delirium continued to experi-
ence cognitive impairment, compared to only 4.4% of the
nondelirious subjects (Bickel et  al., 2008). Similarly, a pro-
spective, matched, controlled cohort study of elderly hip sur-
gery patients demonstrated that the risk of dementia or mild
cognitive impairment (MCI) over a 30-month follow-up was
almost twice as high in patients with postoperative delirium
as in those without (Kat et al., 2008).
A large systematic evidence review demonstrated that the
proportion of persistent delirium in older hospital patients at
discharge, 1, 3, and 6 months was 44.7%, 32.8%, 25.6%, and
21% respectively; one study reported a proportion of 41% at
12 months after discharge (Cole, Ciampi, Belzile, & Zhong,
2009)  (see Figure 41.1). The data suggest that many older
hospital patients do not recover from delirium and that the
persistence of delirium is associated with adverse outcomes.
In some of the studies, the long-term outcomes (mortality,
nursing home placement, cognition, function) of patients
with persistent delirium were consistently worse than the out-
comes of patients who had recovered from delirium.
Furthermore, the data shows that SSD (associated with crit-
ical illness), in the absence of full-blown delirium, has also been
found to result in long-term cognitive dysfunction 2  months
to 6 years following critical illness (e.g., 46%–70% of patients
showed signs of cognitive dysfunction at 1  year and 25% at
6 years) (Cole et al., 2003; Morandi, McCurley, et al., 2012).
On the other hand, the data suggests that among elderly
patients there is a significant acceleration in the slope of cogni-
tive decline in those with Alzheimer’s disease (AD) following
an episode of delirium (Fong et al., 2009). In fact, prospective
data from hospitalized patients with AD (n = 263; median
follow-up duration 3.2  years) found that after adjusting for
dementia severity, comorbidity, and demographic character-
istics, patients who had developed in-hospital delirium expe-
rienced greater cognitive deterioration in the year following
hospitalization relative to patients who had not developed
delirium (Gross et al., 2012). Cognitive deterioration follow-
ing delirium in the year after hospitalization accelerated to
twice the rate of patients who did not develop delirium. This
rate of cognitive deterioration among delirious patients con-
tinued throughout a 5-year period following hospitalization
(Gross et al., 2012). The Vantaa 85+ study (population-based
cohort; n = 553 patients aged ≥85 years at baseline) followed
examined individuals at 3, 5, 8, and 10  years after delirium
and found that delirium increased the risk of incident demen-
tia (OR 8.7, 95% CI, 2.1–3.5) and was associated with the loss
of an additional 1 point per year in the Mini-Mental State
 Discharge 1 month
Rockwood_89 Hodkinson_73
Levkoff_92
Williams_92
Rockwood_93 Jitapunkul_92
Gaudet_93
O’Keefe_97 Rudberg_97
Rudberg_97
Kelly_01
McCusker_03 Kelly_01
Pitkala_04
Lundstrom_05
McAvay_06 Kiely_04
Marcantonio_06
Inouye_07 Marcantonio_06
Combined
Combined* Combined

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Proportion Proportion

3 months 6 months
Levkoff_92 Levkoff_92

Treloar_97 Markup Area

McCusker_03

Kelly_01

Combined Combined
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Proportion Proportion

Individual and combined proportions (and 95% confidence intervals) for persistent delirium at discharge and one, three, and six months
Figure 41.1
after enrollment. Source: (Cole, Ciampi et al., 2009)

Examination compared to those with no history of delirium
(95% CI, 0.11–1.89) (Davis et al., 2012).
A prospective cohort study of 53 patients admitted for
hip fracture surgery aged 75 years and older found that at the
3 months preoperative assessment, patients who had experi-
enced postoperative delirium showed poorer performance
on tests of global cognition and episodic memory, even after
adjusting for age, gender, and baseline cognitive impairment
(Basinski, Alfano, Katon, Syrjala, & Fann, 2010). No differ-
ence was found on tests of attention between patients with
and without history of delirium after adjusting for relevant
confounders. Another study of elderly patients (≥60) under-
going cardiac surgery (i.e., coronary-artery bypass grafting
or valve replacement) found that at 6 months post-surgery, a
higher percentage of patients with delirium than those with-
out delirium had not returned to their preoperative baseline
level of functioning (40% vs. 24%, p = 0.01), although that
difference was not significant at 12  months (31% vs. 20%,
p = 0.055) (Saczynski et al., 2012).
A different type of residual syndrome is post-traumatic
stress disorder (PTSD) secondary to the dramatic and bizarre
delusional and hallucinatory experiences that occur during
a delirious state. The incidence of PTSD related to postop-
erative delirium has been shown to be related to the nature
and level of the sedation and analgesia, degree of factual
memory recall, incidence of delirium, and underlying preva-
lence of preexisting psychiatric morbidity (Blank  & Perry,
1984; Bourgon, 1985; Stukas et al., 1999; Dew et al., 2001;
Jones, Griffiths, Humphris, Skirrow, 2001; Breitbart et  al.,
2002; DiMartini, Dew, Kormos, McCurry, & Fontes, 2007;
Griffiths  & Jones, 2007; Roberts, Rickard, Rajbhandari,  &
Reynolds, 2007; O’Malley, Leonard, Meagher,  & O'Keeffe,
2008; Basinski et al., 2010).
Finally, a number of factors associated with the medical
setting may exacerbate the experience of hypnagogic halluci-
nations, such as medication use, physical restraints, and social
isolation (Jones et  al., 2001). Patients describe memories of
these experiences as being very vivid, detailed, and frighten-
ing. A frightening “hallucinated” memory of a nurse trying
to kill the patient or removing his organs, without a balanc-
ing memory of their medical condition and the care provided
by the same nurse, is likely to contribute to the development
of pseudo-memories with great emotional valence for that
patient.

41. D E L I R I U M • 829
 Studies have found that patients with delirium within the
4 weeks after myeloablative hematopoietic cell transplanta-
tion had significantly more distress and fatigue at 6 months
(P < .004) and at 1 year (P < .03) and had worse symptoms
of depression and post-traumatic stress at 1  year (P < .03),
compared with patients without delirium (Basinski et  al.,
2010). Others have found that family members of delirious
ICU patients can also develop PTSD. Both family caregivers
(76%) and nurses (73%) have reported severe emotional dis-
tress related to a patient’s delirium (Breitbart et al., 2002). In
that study, family members were most distressed by patients’
hyperactivity and functional impairment; nurses were most
distressed by the severity of the patients’ symptoms and by the
presence of perceptual disturbances.
Data suggest that the impact of delirium on outcome may
be changed by therapeutic interventions. (Please see the sec-
tion on delirium management later in this chapter.)
T H E C O S T OF DE L I R I U M
The increased morbidity and extended hospital care associ-
ated with delirium have been associated with greater care
costs (Inouye, 2000; Ely et al., 2004; Milbrandt et al., 2004).
Delirious hospital inpatients require more nursing time per
patient and have higher per-diem hospital costs, in addition to
their increased lengths of stay (Siddiqi, Stockdale, Britton, &
Holmes, 2007). The economic impact of delirium is sub-
stantial, rivaling the health care costs of falls and myocardial
infarction (Hall et al., 1988; Rizzo, McAvay & Tinetti, 1996;
Inouye, 2000).
A retrospective chart review of patients who received
step-down critical care at a large university hospital showed
that the 14% of the ICU population who developed delirium
utilized 22% of total inpatient days (Maldonado, Dhami, &
Wise, 2003). The treated group consisted of the patients for
whom a psychosomatic medicine consult was obtained and
the team implemented at least 80% of the recommended regi-
men (as detailed in the treatment section at the end of this
chapter). The average number of days from symptomatic onset
to resolution was 10.8 days for untreated patients and 6.3 days
for treated patients. As a group, delirious patients were older
(71.3 vs. 63.6  years), remained hospitalized longer (16.4 vs.
6.6  days), and had greater total costs per case ($63,900 vs.
$30,800). Another study of ICU delirium demonstrated that
even after adjusting for age, comorbidity, severity of illness,
degree of organ dysfunction, nosocomial infection, hospital
mortality, and other potential confounders, delirium was
associated with 39% higher ICU and 31% higher hospital
costs (Milbrandt et al., 2004). They also found that the sever-
ity and duration of delirium were associated with incremental
greater care cost. Similarly, a study of hospitalized medically
ill elderly patients the average cost per day was 2.5-fold greater
in those with delirium, and the total excess cost attributable to
delirium ranged from $16,303 to $64,421 per patient (Leslie,
Marcantonio, Zhang, & Leo-Summers, 2008).
The total annual direct healthcare costs attributable to
delirium in the United States might be as high as $152 billion
(Leslie et al., 2008). These costs accrue after hospital discharge
830 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
due to the greater need for long-term care or additional home
health care, rehabilitation services, and informal caregiving.
A recent study looking at costs over one year following an epi-
sode of delirium estimated that delirium is responsible for an
additional cost of $60,000 to $64,000 per patient for the year
following the index hospitalization (Leslie et al., 2008).
E T IOL O G Y OF   DE L I R I U M
Many factors potentially contribute to the development of
delirium (see Table 41.2). I use the mnemonic “End Acute
Brain Failure,” to help recall the 20 of the most clinically rele-
vant risk factors. The reason to use this term is to convey to all
members of the medical team the seriousness of delirium—
indeed, during delirium, the brain is experiencing acute fail-
ure, with potentially disastrous consequences. Only some
of the risk factors are discussed here; readers are referred to
my comprehensive review of delirium risk factors for further
details (Maldonado, 2008b). In general, risk factors can be
grouped as non-modifiable and potentially modifiable.
NON-MODI F I A BL E R I S K FAC TOR S
Recognition of the non-modifiable factors can help identify
patients at high risk for delirium in order to provide enhanced
surveillance and implement preventive measures. (Table 41.3).
Four important non-modifiable factors are older age, baseline
cognitive impairment, severity of underlying medical illness,
and preexisting mental disorders.
Age
Old age is likely to be a contributor due to a greater number
of medical comorbidities and overall frailty. The aging pro-
cess itself is associated with some degree of cognitive deficits
and increased risk of dementia. Finally, aging is associated
with age-related cerebral changes in stress-regulating neu-
rotransmitter and intracellular signal transduction systems.
Chronic neurodegeneration is accompanied by an inflamma-
tory response characterized by a chronic, but selective, acti-
vation of central nervous system (CNS) microglial cells that
are “primed” to produce exaggerated inflammatory responses
to immunological challenges. This inflammation is assumed
to contribute to disease progression through the produc-
tion of inflammatory mediators, including cytokines and
acute phase proteins (Cunningham, Wilcockson, Campion,
Lunnon, & Perry, 2005,; Cunningham et al., 2009). Aging
is also associated with decreased volume of acetylcholine
(Ach)-producing cells and decreased cerebral oxidative
metabolism. Thus, the decline in cognitive functioning asso-
ciated with the normal aging process is aggravated by the
presence of even mild hypoxia, which further inhibits ACh
synthesis and its release.
Among elderly patients in the ICU, the probability of
developing delirium increases by 2% per year of age for each
year after age 65 (Figure 41.2) (Pandharipande et al., 2006).
Many other studies confirm that older age is an independent
Table 41.2 DELIR IUM: PR EDISPOSING AND PR ECIPITATING R ISK FACTORS—“END ACUTE BR AIN FAILUR E”
R ISK FACTOR
Electrolyte imbalance and
dehydration
Neurological disorder and injury
Deficiencies (nutritional)
Age
Cognition
U-Tox (intoxication and withdrawal)
Trauma
Toxins
Endocrine disturbance
Behavioral—psychiatric
Rx and other toxins
Anemia, anoxia, hypoxia, and low
perfusion states
Infectious
Noxious stimuli (pain)
Failure (organ)
APACHE score (severity of illness)
Intracranial processes
Light, sleep, and circadian rhythm
EX A MPLES
Electrolyte disturbances (e.g., hyperammonemia, hypercalcemia, hypo/hyperkalemia,
hypomagnesemia, hypo/hypernatremia)
All neurological disorders: e.g., CNS malignancies, abscesses, cerebrovascular accident (CVA),
vasculitis, multiple sclerosis (MS), epilepsy, Parkinson’s disease, normal pressure hydrocephalus
(NPH), traumatic brain injury (TBI), diffuse axonal injury (DAI), limbic encephalitis (both
non-paraneoplastic and paraneoplastic syndrome).
Of the various forms of sensory impairment, only visual impairment has been shown to contribute
to delirium. Visual impairment can increase the risk of delirium 3.5-fold.
Nutritional deficiencies (e.g., malnutrition, low serum protein/albumin, low caloric intake,
“failure to thrive”), malabsorption disorders (e.g., celiac disease), and hypovitaminosis; specifically
deficiencies in cobalamine (B12), folate (B9), niacin (B3; leading to pellagra), thiamine (B1; leading to
beriberi and Wernicke’s disorder).
Age [>65] and Gender [m>f]
Baseline cognitive functioning, including dementia and other cognitive disorders and/or a past
history of delirium have all been shown to increase the likelihood of delirium.
Substances of abuse—Acute illicit substance intoxication (e.g., cocaine, PCP, LSD, hallucinogens),
as well as poisons, pesticides, solvents, and heavy metals (i.e., lead, manganese, mercury)—and
substances withdrawal.
Physical trauma and injury; heat stroke, hyperthermia, hypothermia, severe burns, including
trauma of surgical procedures.
Various toxins, including bio-toxins (animal poison); heavy metals (lead, manganese, mercury);
insecticides; poisons; carbon dioxide; toxic effect of pharmacological agents (i.e., serotonin
syndrome, neuroleptic malignant syndrome, anticholinergic states); blood levels (toxic levels of
various therapeutic substances; e.g., lithium, VPA, carbamazepine, immunosuppressant agents).
Endocrinopathies such as hyper/hypo-adrenal corticoid; hyper/hypoglycemia; hyper/
hypothyroidism.
Certain psychiatric diagnoses, including undue emotional distress; a history of alcohol and other
substance abuse, as well as depression, schizophrenia, and bipolar disorder have been associated
with a higher incidence of delirium.
Several pharmacological agents have been identified, especially those with high anticholinergic
activity, including prescribed agents (especially narcotics and GABA-ergic agents) and various OTC
agents (especially anticholinergic substances; polypharmacy).
Any state that may contribute to decreased oxygenation (e.g., pulmonary or cardiac failure,
hypotension, anemia, hypoperfusion, intraoperative complications, hypoxia, anoxia, carbon
monoxide poisoning, shock).
Pneumonia, urinary tract infections, sepsis, encephalitis, meningitis, HIV/AIDS.
Data suggest that both pain and medications used for the treatment of pain have been associated
with the development of delirium. Studies have demonstrated that the presence of postoperative
pain is an independent predictor of delirium after surgery. On the other hand, the use of opioid
agents has also been implicated in the development of delirium.
Organ and systemic failure; usually cardiac, hepatic, pulmonary, and renal failure.
Evidence shows that the probability of transitioning to delirium increases dramatically for each
additional point in the Acute Physiology and Chronic Health Evaluation (APACHE II) severity of
illness score.
Stroke (especially non-dominant hemispheric); intracranial bleed; meningitis; encephalitis;
neoplasms.
Sleep deprivation and insomnia, sleep disorders (e.g., obstructive sleep apnea) and disturbances/
reversal in sleep-wake cycle.
(continued)
Table 41.2 (CONTINUED)

R ISK FACTOR EX A MPLES

Uremia and other metabolic Acidosis, alkalosis, hyperammonemia, hypersensitivity reactions; glucose, acid–base disturbances.
disorders

Restraints and any factors causing The use of restraints, including endotracheal tubes (ventilator), soft and leather restraints,
immobility intravenous lines, bladder catheters, and intermittent pneumatic leg-compression devices, casts, and
traction devices all have been associated with an increased incidence of delirium.
Emergence delirium Emergence from medication-induced sedation, coma, or paralysis; which may be associated with
CNS depressant withdrawal, opioid withdrawal, REM-rebound, sleep deprivation.

risk factor among medically ill and surgical patients, with the
increase in rate per year dependent on the specific context in
which incidence is measured (Milstein, Pollack, Kleinman, &
Barak, 2002; Khurana, Gambhir, & Kishore, 2011; Lee et al.,
2011; Rudolph et al., 2011; Srinonprasert et al., 2011).
Baseline Cognitive Impairment
Baseline cognitive deficits, even those not rising to the level
of dementia, significantly increase the risk of developing
delirium. It is well established that individuals with compro-
mised cognitive ability preoperatively (e.g., dementia) are at
greater risk of delirium (Franco et  al., 2010; Inouye, 1998;
Litaker et al., 2001; McNicoll et al., 2003; Benoit et al., 2005).
But recent evidence suggests that decrements in higher order
cognitive functions, such as executive function (e.g., problem
solving, processing speed, planning, complex sequencing,
and reasoning), may also predict postoperative delirium in
the absence of frank cognitive impairment (Rudolph et  al.,
Table 41.3 DELIR IUM R ISK FACTORS
NON-MODIFI ABLE
MODIFI ABLE FACTORS FACTORS
2006). A  study of nondemented elderly patients undergo-
ing elective orthopedic surgery demonstrated that subtle
preoperative attention deficits, as tested by digit vigilance
and reaction time testing, were closely associated with post-
operative delirium (Lowery, & Ballard, 2007). In fact, these
subtle changes predicted a 4-fold to 5-fold increased risk of
postoperative delirium for subjects >1 SD above the sample
means on these variables. Of note, in this study none of the
previously identified risk factors for delirium was significantly
associated with delirium (i.e., urea/creatinine ratio, visual
impairment, blood pressure, sodium levels, hematocrit, bili-
rubin). Finally, a study of nearly 1,000 surgical patients found
that preoperative executive dysfunction was associated with
a greater incidence of postoperative delirium, independent of
other risk factors (Smith et al., 2009). Furthermore, patients
exhibiting both executive dysfunction and clinically signifi-
cant levels of depression were at greatest risk for developing
delirium postoperatively.
A study of elderly patients undergoing hip surgery found that
preoperative Mini-Mental State Exam (MMSE) scores were an
independent predictor of postoperative delirium (Kalisvaart
et al., 2006). Similarly, a study of elderly subjects undergoing
hip or knee replacement demonstrated that the presence of

• Various pharmacological agents, • Older age

especially GABA-ergic and • Baseline cognitive
opioid agents, and medications impairment 0.85
with anticholinergic effects • Severity of underlying
0.80
Probability of Transitioning

• Prolonged and/or uninterrupted medical illness

sedation • Preexisting mental 0.75
Immobility disorders
to Delirium

•
• Acute substance intoxication 0.70
• Substance withdrawal states
• Use of physical restraints 0.65
• Water and electrolyte imbalances
• Nutritional deficiencies 0.60
• Metabolic disturbances and p = 0.004
endocrinopathies (primarily 0.55
deficiency or excess of cortisol) 30 40 50 60 70 80
• Poor oxygenation states (e.g.,
hypo-perfusion, hypoxemia, Age (years)
anemia)
• Disruption of the Age and the probability of transitioning to delirium. The most
Figure 41.2

sleep-wake cycle notable finding related to age was that probability of transitioning

• Uncontrolled pain to delirium increased dramatically for each year of life after 65 years.
• Emergence delirium Adjusted Odds Ratio –1.01 (1.00, 1.02) p = 0.03. Abbreviations: Y-axis =
probability; X-axis = age in years. Source: (Pandharipande et al., 2006)

832 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
dementia increased the incidence of postoperative delirium
from 32% to 100% (Wacker et al., 2006). Approximately 70%
of elderly patients admitted to a specialized delirium ward had
a preexisting cognitive disorder—either dementia or MCI
(Wahlund & Bjorlin, 1999). A recent study of elderly patients
undergoing urological surgery confirmed that age, baseline
cognitive impairment, and previous history of delirium all
were associated with a higher rate of postoperative delirium
(Tognoni et al., 2011). The relationship between cognitive defi-
cits and dementia seems to be reciprocal. So, while it is clear
from the evidence presented above that the presence of baseline
cognitive deficits, including dementia, lowers the threshold to
develop delirium, data suggest that among elderly patients there
is a significant acceleration in the slope of cognitive decline in
patients with AD following an episode of delirium (Fong et al.,
2009). This raises the question of whether prevention of delir-
ium might ameliorate or delay cognitive decline in patients at
risk, particularly those with AD.
Severity of Medical Illness
Studies have also shown that the severity of the patient’s
underlying medical problems has a significant effect on the
development and progression of delirium (Ouimet et al.,
2007; Girard et al., 2008; Maldonado, 2008b; Khan, Kahn,
& Bourgeois, 2009; Pisani, Kong, et al., 2009; Maldonado,
2011; Srinonprasert et al., 2011; Ryan et al., 2013). A study
of adult, mechanically ventilated ICU patients found that
increased severity of illness, as measured by the modified
Acute Physiology and Chronic Health Evaluation (APACHE
II; a modified-APACHE II is reached by subtracting the
Glasgow Coma Scale score from the usual calculation of the
APACHE II total score) was associated with a greater prob-
ability of developing delirium (Pandharipande et al., 2006).
The incremental risk increased with APACHE score until the
latter reached 18 (see Figure 41.3), suggesting that for each
additional point on the APACHE II scale, the probability of
delirium increased by 6%. These findings were replicated in a
0.70
Probability of Transitioning
0.65
to Delirium
0.60
0.55
0.50
0.45
10 15 20
APACHE II Score
Severity of Illness and the Probability of Transitioning to
Figure 41.3
Delirium. The probability of transitioning to delirium increased dra-
matically for each additional point in the Acute Physiology and Chronic
Health Evaluation II (APACHE II) severity of illness score, until reach-
ing a plateau APACHE score of 18. (Pandharipande et al., 2006)
study of elderly patients undergoing hip surgery (Kalisvaart
et  al., 2006) and another, of critically ill patients (Tsuruta
et al., 2010), where APACHE II scores were identified as an
independent predictor of delirium when controlling for the
use of mechanical ventilation and length of ICU stay.
Preexisting Mental Disorders
A 1996 study retrospectively investigated incidence rates
and risk factors for delirium among hospitalized psychiat-
ric patients (Ritchie, Steiner,  & Abrahamowicz, 1996). The
overall incidence of delirium in the study sample was 15%.
Patients with bipolar disorder had the highest incidence of
delirium (36%). Only 48% of delirious patients were actually
recognized as having delirium at the time it occurred. The
hospital stays of patients with delirium were 62% longer than
those without.
Delirious mania, also known as “Bell’s mania,” is defined
as “a syndrome of acute onset of excitement, grandiosity,
emotional lability, delusions and insomnia characteristic of
mania, and the disorientation and altered consciousness char-
acteristic of delirium. Almost all patients exhibit signs of cata-
tonia” (Bell, 1849; Fink, 1999). The syndrome consists of a
constellation of symptoms that can arise from both psychotic
and affective psychiatric diseases, as well as from many medi-
cal diseases. It has been described as a subtype of the catatonia
syndrome, having a rapid onset with signs of mania, delirium,
and catatonia (Friedman, Mufson, Eisenberg, & Patel, 2003).
Among patients undergoing cardiac surgery—after adjust-
ment for sex, older age, cross-clamp time, hemoglobin, and
psychotropic drug use—major depression was significantly
associated with delirium (OR  =  3.86, 95% CI 1.42–10.52,
p = 0.001) (Tully, Baker, Winefield, & Turnbull, 2010). A more
recent study of cardiac surgery patients found that patients
with a preoperative diagnosis of depression had an indepen-
dently associated increased risk of developing postsurgical
delirium (Kazmierski, Banys, Latek, Bourke,  & Jaszewski,
2013). Thus, major depressive disorder has been associated
with incident delirium after various types of surgical proce-
dures, as well as among general hospital inpatients, indepen-
dent of other risk factors (Schneider et al., 2002; Dasgupta &
Dumbrell, 2006; Kazmierski et  al., 2006; McAvay et  al.,
2007; Smith et al., 2009). A study of geriatric surgical subjects
showed that patients with a greater number of preoperative
depressive symptoms were more likely to develop postop-
erative delirium and experience a longer duration of postop-
erative delirium, even after adjusting for covariates (i.e., age,
educational level, functional status, and preoperative alcohol
use) (Leung, Sands, Mullen, Wang, & Vaurio, 2005).
p = 0.004
Long-term benzodiazepine use is usually associated with
the presence of an underlying psychiatric disorder (e.g., anxiety,
25 30 depression, insomnia, alcohol abuse, benzodiazepine abuse), and
thus it may be considered a marker of underlying psychiatric
comorbidity. Studies have looked at the presence of postoperative
delirium between long-term benzodiazepine users and nonusers
and found that the incidence of delirium after orthopedic sur-
gery in the elderly was significantly more frequent in long-term
benzodiazepine users (26%) than in short-term users (13%)
41. D E L I R I U M • 833
(Kudoh,Takase, Takahira, & Takazawa, 2004). Benzodiazepines
have been demonstrated to have a negative effect on memory
(Lister, 1985; Ghoneim  & Mewaldt, 1990; Ghoneim, Block,
Ping, el-Zahaby, & Hinrichs, 1993), and it has been demonstrated
that benzodiazepine use induces both nonamnestic and amnes-
tic MCI (Tannenbaum, Paquette, Hilmer, Holroyd-Leduc,  &
Carnahan, 2012). Furthermore, chronic BZDP users had a sig-
nificantly higher risk of cognitive decline in global cognitive and
attention tests compared with non-BZDP users (Mura et  al.,
2013). It appears that even though there may be an improvement
after BZDP discontinuation, there remains a significant impair-
ment in most areas of cognition when compared to controls
(Barker, Greenwood, Jackson, & Crowe, 2004, 2005).
As we discussed before, there is a relationship between
dementia and development of delirium. New data sug-
gest that the use of benzodiazepine is associated with an
increased risk of Alzheimer’s disease (adjusted OR 1.51, 95%
CI, 1.36–1.69) (Billioti de Gage et  al., 2014). Furthermore,
the risk increases with cumulative exposure:  1.32 (1.01 to
1.74) for 3 to 6  months, and 1.84 (1.62 to 2.08) for more
than 6 months (Billioti de Gage et al., 2014). Similarly, in a
large nested-control study, after controlling for multiple vari-
ables (i.e., age, gender, education level, living alone, alcohol
consumption, psychiatric history, and depressive symptom-
atology), the ongoing use of benzodiazepines was associated
with a significantly increased risk of dementia (adjusted OR,
1.7; 95% CI, 1.2–2.4), while former use was associated with
a significantly increased risk of dementia (adjusted OR, 2.3;
95% CI,1.2–4.5) (Lagnaoui et al., 2002). Others have found
similar results, with a marked increased incidence of dementia
(OR = 3.50, 95% CI 1.57 to 7.79, p = 0.002) among those on
long-term BZDP use (Gallacher et al., 2012). Finally, among
subjects seeking acute treatment for alcohol withdrawal, 6.9%
developed alcohol withdrawal delirium despite adequate
treatment with benzodiazepine agents (Palmstierna, 2001).
These findings raise the possibility that the use of benzodiaze-
pines, when combined with other risk factors, may contribute
or accelerate the rate of cognitive decline following delirium.
P OT E N T I A L LY MODI F I A BL E R I S K FAC TOR S
Among the significant factors that are amenable to early
intervention or prevention are electrolyte imbalances and
nutritional deficiencies, metabolic disturbances and endo-
crinopathies, low oxygenation states, environmental factors
disrupting the sleep-wake cycle and impeding adequate rest,
pain and its treatment, the use of drugs with anticholinergic
effects, substance intoxication and withdrawal states, the use
of physical restraints, and the development of “emergence
delirium” (i.e., an altered mental status occurring as a patient
“emerges” from deep sedation or medication-induced coma).
Electrolyte Imbalances and Nutritional Deficiencies
Irrespective of the etiology, a water and electrolyte imbalance
(primarily magnesium, phosphate, potassium, and chloride)
provoking a hypo- or hyperosmolar state causes metabolic
encephalopathy or delirium (Mattle, 1985; Schmickaly et al.,
834 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
1989; Wetterling, Kanitz, Veltrup, & Driessen, 1994; Aldemir
et  al., 2001; Lawlor, 2002; Sagawa, Akechi, et  al., 2009;
Caplan & Chang, 2010). The correction of these electrolyte dis-
turbances has been shown to significantly shorten the duration
of delirium (Koizumi, Shiraishi, Ofuku, & Suzuki, 1988). The
relationship between nutritional deficiencies and delirium has
long been documented, and it is the clearest in patients suffering
from Wernicke’s encephalopathy (B1 or thiamine deficiency)
(Hoes, 1979; Newman, Grocott, et al., 2001; Onishi, Sugimasa,
Kawanishi, & Onose, 2005). In addition, vitamin B6 and B12
deficiency have all been associated with delirium (Peters  &
Neumann, 1960; Hoes, 1979; Buchman, Mendelsson, et  al.,
1999; Lerner & Kanevsky, 2002). Some data suggest that vita-
min B12 supplementation (in patients with documented defi-
ciencies) reduced the duration of delirium in elderly demented
subjects. And, others have found a relationship between low
albumin levels and the development of delirium among elderly
demented patients (Culp & Cacchione, 2008).
Metabolic Disturbances and Endocrinopathies
Metabolic disturbances and various endocrinopathies have
been associated with the development of delirium, primar-
ily deficiency or excess of cortisol (Thiele  & Hohmann,
1961; Millerowa, 1966; McIntosh et al., 1985; Basavaraju &
Phillips, 1989; Olsson, Astrom, Eriksson,  & Forssell, 1989;
Fassbender, Schmidt, Mossner, Daffertshofer,  & Hennerici,
1994; O’Keeffe & Devlin, 1994; Johansson, Olsson, Carlberg,
Karlsson,  & Fagerlund, 1997; Olsson, 1999; Robertsson
et al., 2001; Abildstrom, Christiansen, Sirsma, & Rasmussen,
2004; Nemoto, Kawanishi, Suzuki, Mizukami,  & Asada,
2007; Weng, Chang,  & Weng, 2008)  or thyroid hormone
(Vidal  & Vidal, 1961; Goldfarb, Varma,  & Roginsky, 1980;
Nibuya, Suda, Mori, & Ishiguro, 2000; El-Kaissi, Kotowicz,
Berk, & Wall, 2005), and either high or low blood glucose lev-
els (Fishbain & Rotundo, 1988; Aldemir, Ozen, et al., 2001;
Miller, 2008). There have been case reports of Cushing’s
disease caused by adrenocorticotropic hormone (ACTH)–
producing pituitary adenomas, in patients who lacked the
phenotypical signs of hypercortisolism (e.g., facial plethora,
supraclavicular fat pads, buffalo hump, truncal obesity, and
purple striae) and whose most prominent and distressing
symptoms were severe myopathy and altered mental status
(Tran  & Elias, 2003). Similarly, there have been reports of
hypoactive delirium as the primary presentation of a patient
with tonsillar abscess and associated panhypopituitarism
and secondary hyponatremia (Umekawa Yoshida, Sakane, &
Kondo, 1996)  and of a patient experiencing delirium sec-
ondary to serendipitously diagnosed hypocortisolemia, who
presented none of the usual signs of adrenal insufficiency or
endocrinopathy (Fang & Jaspan, 1989). Others have demon-
strated that, among demented inpatients with no acute medi-
cal illness, there was a strong relationship between delirium
and dexamethasone suppression test (DST) pathology, irre-
spective of age and the severity of dementia; suggesting the
possibility that an impaired hypothalamus–pituitary–adrenal
(HPA) system and a low delirium threshold may lead to delir-
ium as a response to stress (Robertsson et al., 2001).
 Low Oxygenation States Premorbid Hemoglobin
12
Poor oxygenation (i.e., hypoperfusion, hypoxemia, anemia)
has long been associated with the development of delirium
(Kinney, Duke, et al., 1970; Weissman et al., 1984; McGee,
Veremakis, & Wilson, 1988; Fine, Anderson, Rothstein, 11
Williams, & Gochuico, 1997; Maldonado, 2008a,b).

gm/dl
Inadequate oxidative metabolism may be one of the under-
lying causes of the basic metabolic problems initiating the
10
cascade that leads to the development of delirium; namely,
inability to maintain ionic gradients, causing cortical spread-
ing depression (i.e., spreading of a self-propagating wave of
cellular depolarization in the cerebral cortex) (Moghaddam, 9
Schenk, Stewart, & Hansen, 1987; Somjen et al., 1989, 1990; Delirium Non-delirium
Iijima, Shimase, Iwao, & Sankawa, 1998; Shimizu-Sasamata, (N = 30) (N = 70)

Bosque-Hamilton, Huang, Moskowitz, & Lo, 1998; Basarsky, Note: p<0.004; Student’s t-test with equal variances not assumed
(Levene’s test).
Feighan, & MacVicar, 1999); abnormal neurotransmit-
ter synthesis, metabolism, and release (Globus, Busto et al., Premorbid Hematocrit
1988b, 1989, 1990; Busto et al., 1989; 1990; Globus, Wester, 36
Busto, & Dietrich, 1992; Takagi, Ginsberg, et al., 1994; Busto,
Dietrich, Globus, Alonso, & Dietrich, 1995; Busto, Dietrich,
et al., 1997); and a failure to effectively eliminate neurotoxic
34
byproducts (Busto, Globus, et al., 1989; Globus, Busto, et al.,
1990; Globus, Alonso, et al., 1995).
Percent
Study findings correlate the development of delirium
with findings suggestive of premorbid oxidative metabo- 32
lism impairment. In a study of ICU patients, three measures
of oxygenation (i.e., hemoglobin, hematocrit, and pulse
oximetry) were worse in the patients who later developed
30
delirium (Figure 41.4) (Seaman, Schillerstrom, Carroll,  &
Delirium Non-delirium
Brown, 2006). Similarly, two measures of oxidative stress (N = 30) (N = 70)
(sepsis, pneumonia) occurred more frequently among those Note: p<0.006; Student’s t-test with equal variances not assumed.
diagnosed with delirium. Others have demonstrated that
intraoperative cerebral oxygen desaturation (as measured Premorbid Pulse-Oximetry
by cerebral oximeter) is significantly associated with an 94
increased risk of postoperative delirium and early cognitive
decline and was associated with a nearly threefold increased
risk of prolonged hospital stay after cardiac surgery (CABG)
(Figure 41.5) (Slater et al., 2009). Finally, another large study 91
Percent Saturation

of urgent cardiac surgery demonstrated that preoperative

and intraoperative ScO2 readings were lower in the group
of patients who developed delirium. The binary logistic
regression identified older age, lower MMSE, neurological 88
or psychiatric disease, and lower preoperative ScO2 as inde-
pendent predictors of postoperative delirium (Figure 41.6)
(Schoen et al., 2011).
85
Delirium Non-delirium
(N = 30) (N = 70)
Sleep Disturbances Note: p<0.006; Student’s t-test with equal variances not assumed.

Sleep is another factor that seems to play a significant role in
developing delirium in the ICU. Sleep deprivation has long
been linked to the development of delirium (Lipowski, 1987)
and psychosis (Berger, Vollmann, et al., 1997). Patients in the
ICU experience severe alterations of sleep, with sleep loss,
sleep fragmentation, and sleep–wake cycle disorganization.
Continuous polysomnographic recordings have demonstrated
that despite sedative practices, the average amount of sleep in
ICU patients is limited (e.g., as short as 1 hour and 51 minutes
Oxidative Stress Hypothesis. Three measures of oxygenation
Figure 41.4
(Hg, Hct, pulse ox.) were worse in the patients who later developed
delirium.
per 24-hour period) (Aurell & Elmqvist, 1985). Sleep in ICU
patients is characterized by prolonged sleep latencies, sleep
fragmentation, decreased sleep efficiency, frequent arousals, a
predominance of stage 1 and stage 2 non-rapid eye movement
(REM) sleep, decreased or absent stage 3 and 4 non-REM sleep,

41. D E L I R I U M • 835
 changes/placement), noise, light, and discomfort (Weinhouse
8 & Schwab 2006; Drouot et al., 2008; Friese, 2008b). Another
7 large factor in sleep debt in ventilated patients is continuous
sedation. In fact, studies have demonstrated that propofol and
6 benzodiazepine agents (the primary agents used for continu-
5 ous sedation) have a negative effect on sleep (e.g., severe altera-
tions of sleep with sleep loss, ↓ REM sleep, loss of circadian
4 rhythm, extreme sleep fragmentation, sleep–wake cycle dis-
3 organization) (Drouot et al., 2008; Matthews, 2011; Boyko,
Ording, & Jennum, 2012; Hofhuis, Langevoort, Rommes, &
2
Spronk, 2012; Kamdar, Needham, & Collop, 2012; McKinley
1 et al., 2012; Nakos, 2012; Watson, Ceriana, & Fanfulla, 2012;
Andersen, Boesen, & Olsen, 2013; Brummel & Girard, 2013;
0
Gomez Sanz, 2013).
No Cerebral Cerebral
Desaturation Desaturation

Cerebral Desaturation and Postoperative Delirium. Intra-

Figure 41.5
operative cerebral O2 desaturation was a significant risk factor for
postoperative delirium in CABG patients. (Slater et al., 2009)
and decreased or absent REM sleep (Weinhouse & Schwab,
2006; Drouot, Cabello, d’Ortho, & Brochard, 2008; Friese,
2008b).
There are many factors that may contribute to sleep debt
in the hospital, and particularly in the ICU. These include fre-
quent diagnostic and therapeutic interventions and procedures;
anxiety, fear, pain, and underlying illness and disease severity;
and environmental factors, such as around-the-clock medi-
cal care (e.g., suctioning, ventilator settings, blood draws, line
30
no Delirium
25
Delirium
p = 0.024
20
Delta ScO2base
Pain Management
As in the case with sleep, both the experience of pain and some
of the pharmacological agents used for the treatment of pain
have been associated with the development of delirium. In fact,
the presence of postoperative pain has been shown to be an
independent predictor of postoperative delirium (Vaurio et al.,
2006). There is also a direct relationship between levels of preop-
erative pain and the risk of developing postoperative delirium.
Unfortunately, some opioid agents have also been implicated
in the development of delirium (Fong, Sands, & Leung, 2006;
Vella-Brincat & Macleod, 2007; Wang, Sands, Vaurio, Mullen,
& Leung, 2007). In fact, opioid agents have been implicated in
nearly 60% of the cases of delirium in patients with advanced
cancer (Centeno, Sanz, & Bruera, 2004). Studies among cancer
patients revealed significant associations between opioids and
delirium (Gaudreau et al., 2007). The association remained
significant after adjustment for corticosteroid, benzodiazepine,
and antipsychotic exposure using generalized estimating equa-
tion regressions (OR of 1.37; P = .0033). Others have suggested
that patients who used oral opioid analgesic agents as their
sole means of postoperative pain control are at decreased risk
of developing delirium compared to subjects using intravenous

(IV) opioid-based, patient-controlled analgesia (PCA) (Vaurio

15
10
5 agement and lower delirium rating scores (Morita et al., 2005).
0 linergic potential have all been implicated in opioid-induced
cerebral oxygen saturation ScO2base>59.5%
(ScO2) base<59.5%
Preoperative and Intraoperative Cerebral O2 Saturation
Figure 41.6
(ScO2) Values Influenced the Risk of Post-op Delirium Intraoperative
changes in ScO2 in patients with or without delirium classified by
normal or low preoperative ScO2. Delta ScO2base, difference between
preoperative regional cerebral oxygen saturation with oxygen supple-
mentation and minimal intraoperative regional cerebral oxygen satura-
tion; ScO2base, regional cerebral oxygen saturation with supplemental
oxygen. (Schoen et al., 2011)
et al., 2006; Wang et al., 2007).
Some opioid agents (e.g., meperidine) may have greater
deliriogenic potential than others (Marcantonio et al., 1994;
Morrison et al., 2003; Fong et al., 2006). Studies suggest that
an opioid rotation to less deliriogenic agents like fentanyl or
hydromorphone has been associated with improved pain man-
The exact mechanism of delirium causation is unclear, but
agent half-life, the presence of active metabolites, and anticho-
delirium (Slatkin & Rhiner, 2004). Some clinicians prefer the
use of PCA, as the common wisdom is that patients usually
“self-titrate” and end up receiving lower doses of medication,
thus lowering delirium risk. This may be true in the cognitively
intact individual; but in the cognitively impaired or already
delirious patient, PCAs are contraindicated. The risk of a PCA
apparatus for a patient who is already confused is that they may
not know how to use it, what to do with the button. Essentially
the question to answer is, is the patient cognitively intact and

836 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
competent enough to use PCA appropriately? A simple rule: if Table 41.4 COMMONLY USED MEDICINES THAT
the patient is experiencing any sign of confusion (e.g., disorien- HAVE ANTICHOLINERGIC EFFECTS
tation, finger agnosia), do not use a PCA. This may be resumed
ANTIHISTA MINES
when the patient’s delirium resolves.
Chlorpheniramine
Medications Cyproheptadine
Besides opioid agents, a great number of medications have Dexchlorpheniramine
been associated with an increased risk of delirium (Table
41.4). Some have said that the highest incidence of medica- Diphenhydramine
tion-induced delirium has been observed in patients tak- Dimenhydrinate
ing the greater number of agents (Inouye & Charpentier,
1996), agents with high anticholinergic potential (Tune, Carr, Hydroxyzine
Cooper, Klug, & Golinger, 1993), and medications with signif- Meclizine
icant psychoactive effects (Francis, Martin, & Kapoor, 1990;
Olofsson, Weitzner, Valentine, Baile, & Meyers, 1996; Brauer, Promethazine
Morrison, Silberzweig, & Siu, 2000; Lawlor et al., 2000; ANTIPAR K INSONI AN
Tuma & DeAngelis, 2000; Morita et al., 2001; Breitbart et al.,
2002; Gaudreau & Gagnon, 2005). Among this latter group, Benztropine
opioids, corticosteroids, and benzodiazepines have been iden- Biperiden
tified as major contributors to delirium in several studies (see
Figure 41.7) (Gaudreau, Gagnon, et al., 2005b). In fact, nar- Levodopa
cotics, benzodiazepines, corticosteroids, and other psycho- Trihexyphenidyl
active agents use have been associated with 3-fold to 11-fold
increases in the prevalence of delirium (Morita et al., 2001; CAR DIOVASCULAR

Morrison et al., 2003; Gaudreau & Gagnon, 2005; Gaudreau

Amiodarone
et al., 2005b, 2007; Fong et al., 2006; Pisani, Murphy, et al.,
2009; van Munster et al., 2010). Beta-blockers
There is significant evidence linking a medication’s anticho-
Captopril
linergic potential (as measured by serum anticholinergic activ-
ity) and their incidence of causing delirium (Tune, et al., 1981, Chlorthalidone
1993; Tune, Carr, Hoag, & Cooper, 1992; Golinger al., 1987;
Digoxin
Milusheva et  al., 1990; Inouye  & Charpentier, 1996; Flacker
et al., 1998; Tune & Egeli, 1999; Tune, 2000; Kojima, Terao, & Diltiazem
Yoshimura, 1993; Meyer, Meyer,  & Kressig, 2010). Several
Dipyridamole
studies have identified some of these substances, and we have
become increasingly aware of the importance of understand- Disopyramide
ing the cumulative serum anticholinergic activity (Tune et al.,
Dopamine
1981; Tune & Folstein 1986; Tune et al., 1992, 1993; Tune &
Egeli, 1999; Tune, 2000) (see Table 41.5). Others have demon- Ergotamine
strated that, among medical inpatients, exposure to anticholin-
Furosemide
ergic agents is an independent risk factor for the development
of delirium and is specifically associated with a subsequent Hydrochlorothiazide
increase in delirium symptom severity (Han et al., 2001).
Hydralazine
Despite its apparent usefulness, there are limitations to
the use of the serum anticholinergic activity (SAA) assay in Isosorbide mononitrate
clinical practice: (1) assays of standard drug solutions do not
Lanoxin
account for pharmacokinetic differences among drugs, which
limits the interpretation of such measurements; (2)  emerg- Lidocaine
ing evidence has suggested that anticholinergic medications
Methyldopa
may not be the only cause of elevated SAA (Carnahan, Lund
et  al., 2002). For example, naturally occurring SAAs have Nifedipine
been described in the absence of exposure to known anti-
Procainamide
cholinergic pharmacological agents; their etiology, although
they are postulated to be related to the primary illness state Quinidine
(or aging process), and their clinical significance remain Triamterene
unknown (Flacker & Wei, 2001). Despite these limitations,
elevated SAA has been consistently associated with cognitive Warfarin

(contiuned)
41. D E L I R I U M • 837
Table 41.4 (CONTINUED) Table 41.4 (CONTINUED)
CENTR AL NERVOUS SYSTEM
Propantheline
Barbiturates Scopolamine
Benzodiazepines GENITOUR INARY

Carbamazepine Oxybutinin
Chloral hydrate IMMUNE ACTIVITY/IMMUNOSUPPR ESSION

Dextromethorphan
Azathioprine
Heterocyclic antidepressants (esp. amitriptyline, clomipramine,
Cyclosporine
doxepin, imipramine, trimipramine)
Interleukin-2
Lithium
Interferon
Low-potency neuroleptics (esp. chlorpromazine,
thioridazine) INFECTIOUS

Midazolam Amphotericin-B
Meprobamate Ampicillin
Monoamino oxidase inhibitors (MAOIs) Cefalothin
Phenytoin Cefamandole
Propofol Cefoxitin
SGAs (some: clozapine, olanzapine) Clindamycin
SSRIs (some: paroxetine) Cotrimazole
CORTICOSTEROIDS
Cycloserine
Betamethasone Fluoroquinolones
Cortisone Gentamicin
Dexamethasone Itraconazole
Hydrocortisone Mefloquine
Prednisolone Piperacillin
GASTROINTESTINAL
Tobramycin
Atropine Vancomycin
Belladonna MUSCLE R ELA X ANTS

Clinidum Baclofen
Dicyclomine Carisoprodol
Dimenhydrinate Cyclobenzaprine
Diphenoxylate Metaxalone
H2-antagonists (cimetidine, famotidine, Orphanedrine
nizatidine, ranitidine)
Pancuronium
Hyoscyamine
Tizanidine
Lomotil
ONCOLOGY
Loperamide
Asparaginase
Meclizine
Fluorouracil (5-FU)
Metoclopramide
Tamoxifen
Promethazine
(contiuned)
Table 41.4 (CONTINUED) Table 41.5 ANTICHOLINERGIC DRUG LEVELS IN 25
OTC MEDICATIONS R ANK ED BY THE FR EQUENCY
OF THEIR PR ESCR IPTION FOR ELDER LY PATIENTS
Cold/sinus preparations (antihistamines, pseudoephedrine)
ANTICHOLINERGIC DRUG
Sleep aids (diphenhydramine, alcohol-containing elixirs) LEVEL § (NG/ML OF ATROPINE
MEDICATION† EQUIVALENTS)
Stay-awake preparations
PAIN M ANAGEMENT 1. Furosemide 0.22

2. Digoxin 0.25
Codeine
3. Dyazide 0.08
Meperidine
4. Lanoxin 0.25
NSAIDs (esp. indomethacin and COX2-inhibitors)
5. Hydrochlorothiazide 0.00
Oxycodone
6. Propranolol 0.00
Pentazocine
7. Salicylic acid 0.00
Propoxyphene
8. Dipyridamole 0.11
Tramadol
9. Theophylline anhydrous 0.44
R ESPIR ATORY SYSTEM
10. Nitroglycerin 0.00
Theophylline
11. Insulin 0.00

12. Warfarin 0.12

impairment and delirium in a number of research settings. †= At a 10 M concentration.

-8

In recent years, new data has renewed the concern regard- § =Threshold for delirium = 0.80ng/ml.
ing the relationship between anticholinergic agents and the Source: Tune et al., 1992
neurocognitive disorders, including delirium and demen-
tia. In fact, cumulative evidence has demonstrated a strong
correlation between the use of anticholinergic agents and inci-
dent dementia (Lopez-Alvarez, Zea Sevilla et al. 2014, Gray,
80 Anderson et al. 2015, Mate, Kerr et al. 2015) and an increased
d
risk of hospitalization for confusion (Kalisch Ellett, Pratt
70 et al. 2014).
Among the main offenders, GABA-ergic medications have
60 a been increasingly implicated in the development of delirium
c (Marcantonio et al., 1994; Meuret et al., 2000; Pain, Jeltsch,
50 et  al., 2000; Wang et  al., 2000; Ely, Gautam, et  al., 2001).
Percent of Cases

Several studies have demonstrated a relationship between

40
benzodiazepine use and delirium (Tune  & Bylsma, 1991;
Marcantonio et  al., 1994; Ely, Gautam, et  al., 2001; Kudoh
30 a
c
et al., 2004). In fact, the benzodiazepine agent lorazepam has
b c been found to be an independent risk factor for daily transi-
20 f
tion to delirium (see Figure 41.8) (Pandharipande et al., 2006).
e f
A  recent study found that among ventilated burn patients,
10
e exposure to benzodiazepines was an independent risk factor
for the development of delirium (P <.001) (Agarwal, et  al.,
0
Opioids Cortico- Benzo- 2010). Similarly, the use of various general anesthetics has
steroids diazepines been associated to structural changes leading to disruption
aBreitbart et al. 200226 of blood–brain barrier (BBB) integrity (Forsberg et al., 2014).
bMorita et al. 200129 This will be discussed in further detail under the inflamma-
cTuma and DeAngelis 200030
dLawlor et al. 200014
tory hypothesis of delirium.
eOlofsson et al. 199627 There are many mechanisms by which sedative agents,
fFrancis et al. 19999 particularly GABA-ergic agents, mediate delirium. These
may include interruption of physiological melatonin and
Delirium Potentially Caused by Opioids, Corticosteroids, and
Figure 41.7 sleep patterns, interference with central cholinergic transmis-
Benzodiazepines in Six Case Series. (Gaudreau, Gagnon, et al., 2005b) sion, disruption of thalamic gating, and acute and long-term

41. D E L I R I U M • 839
 1.0
Probability of Transitioning
0.9
to Delirium
0.8
0.7
0.6
0 10 20
Lorazepam Dose (mg)
Lorazepam and the Probability of Transitioning to Delirium
Figure 41.8
The probability of transitioning to delirium increased with the dose
of lorazepam administered in the previous 24 hours. This incremental
risk was large at low doses and plateaued at around 20 mg/day.
Abbreviations: Y-axis = delirium risk; X-axis = lorazepam dose (in mg) (Pandharipande et al., 2006)
cognitive disturb ances. Please see Box 41.2 for a comprehen-
sive list of the 12 mechanisms by which sedative and analgesic
agents mediate their deliriogenic effects. Similar to more clas-
sic anticholinergic agents, acute benzodiazepine use has been
linked to agitation, confusion and delirium (Daman Willems
and Dillon 1986, Short, Forrest et al. 1987, Miller and Gold
1991, O'Reilly and Smith 1991, Trewin, Lawrence et al. 1992,
Lechin, van der Dijs et al. 1996, Hofmann 2013); and chronic
use is linked to Alzheimer’s disease (Billioti de Gage, Moride
et al. 2014, Kmietowicz 2014, Yaffe and Boustani 2014).
Box 41.2 SEDATIVE AND ANALGESIC AGENTS’
DELIRIOGENIC MECHANISMS:
Sedative agents (mostly GABA-ergic) and opioids may contrib-
ute to the development of delirium by one of twelve possible
mechanisms:
1. Acutely, benzodiazepines (BZDP) are associated with the
development of psychomotor retardation, cognitive blunt-
ing, ataxia, poor balance, and decreased mobility, all known
contributing factors for delirium (Sarasin, Ghoneim, &
Block, 1996).
2. BZDP have been demonstrated to have a negative effect on
memory (Lister, 1985; Ghoneim & Mewaldt, 1990; Ghoneim
et al., 1993).
3. BZDP use induces both non-amnestic and amnestic mild
cognitive impairment (Tannenbaum et al., 2012)
4. Chronic administration of BZDP induces a down-regulation
of BZDP-receptors (Hutchinson, Smith, & Darlington,
1996), and a reduction in the number of these receptors
seems to be correlated with cognitive decline (Shimohama,
Taniguchi, Fujiwara, & Kameyama, 1988).
840
5. Long term BZDP use may induce a limitation in cognitive
reserve capacity, which further reduces a person’s ability to
cope with early phase brain lesions by soliciting accessory
neuronal networks (Stern, 2002).
6. New evidence suggests that BZDP use may be associated
with an increased cumulative risk of dementia (Billioti de
Gage et al., 2014).
p = 0.003 7. BZDP use interferes with central cholinergic function
muscarinic transmission at the level of the basal forebrain
30 40 and hippocampus (i.e., causing a centrally mediated acetyl-
choline deficient state) (Schneck & Rupreht, 1989; Meuret,
Backman, et al., 2000; Pain et al., 2000; Wang et al., 2000);
furthermore, BZDP activate GABAA receptors, which
inhibit glutamate effect on NMDA receptors, thus limiting
Ach release (Cervetto & Taccola, 2008).
8. BZDP use interferes with physiological sleep patterns
(e.g., ↓ slow wave sleep→ ↑REM latency→↓REM periods
duration→REM deprivation) (Borbely, Mattmann, et al., 1985;
Achermann & Borbely, 1987; Bastien, LeBlanc, Carrier, &
Morin, 2003, Qureshi & Lee-Chiong, 2004; Mazza et al., 2014).
9. BZDP use disrupts the circadian rhythm of melatonin
release (Olofsson et al., 2004).
10. Long-term use of BZDP increases compensatory
up-regulation of N-methyl D-aspartate (NMDA) and
kainate receptors and Ca2+ channels (Chaudieu, St-Pierre,
Quirion, & Boksa, 1994; Heikkinen, Moykkynen, & Korpi,
2009); enhancing NMDA-induced neuronal damage (Zhu,
Cottrell, & Kass, 1997).
11. BZDP use disrupts thalamic gating function (Gaudreau &
Gagnon, 2005).
12. Long-term BZDP use may lead to CNS-depressant with-
drawal syndromes, upon abrupt discontinuation.
Immobility
Immobility—due to disease conditions, physical restraints,
or medical apparatus that effectively limits the patient’s
mobility (e.g., ET tube [ventilator], IV lines, bladder cath-
eters, intermittent pneumatic leg-compression devices, casts,
traction devices)— has been shown to increase the inci-
dence of delirium (McCusker, Verdon, Caplan, Meldon, &
Jacobs, 2002; Inouye et al., 2007). Some may include chemi-
cal restraints in this group. Data show that patients who are
mobilized early (i.e., alerted and provided physical and occu-
pational therapy) during the course of illness demonstrate
greater return to independent functional status at hospital
discharge (59% vs. 35%), spend less time in the ICU (2 vs.
4  days), and experience less delirium (33% vs. 57%) com-
pared to a control cohort (Schweickert et al., 2009).
Both psychoactive medications and immobility may
contribute to emergence delirium (ED) or emergence agi-
tation (EA), a postoperative behavior that may occur upon
• P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
awakening in patients exposed to general anesthesia (Bastron
& Moyers, 1967; Gutstein, 1996; Haynes, 1999; Fong et al.,
2006; Maldonado, 2008b) and that has been described in
both children and adults (Albin, Bunegin, Massopust, &
Jannetta, 1974; Lepouse, et al., 2006; Abu-Shahwan, 2008;
Fronapfel, 2008; Hudek, 2009). Even though there is some
suggestion that the incidence may be increased after the
use of volatile (inhaled) anesthetic agents (Bajwa, Costi, &
Cyna, 2010), many other agents commonly used for anesthe-
sia and sedation in the ICU have been implicated (as already
discussed above). All forms of conventional sedation may
be associated with emergence delirium, although some are
better than others. As a general rule, propofol is less delirio-
genic than midazolam, but dexmedetomidine is less delirio-
genic than both. The use of opioids as a sedative agent may
be one of the main contributors of emergence delirium and
should be avoided. The rule should be: Use sedative agents
for sedation and narcotics for pain management, rather than
the commonly practiced narcosedation. Another contribu-
tor to emergence delirium may be CNS depressants (e.g.,
narcotics, benzodiazepines, other GABA-ergic agents) when
a patient has been exposed to these agents for a long period
of time, and the primary team feels “it is time to extubate”
so the patient is quickly taken off these agents rather than
being carefully weaned off them. Being more mindful of
this practice and its effects, and better planning, may signifi-
cantly reduce delirium’s occurrence.
N E U ROPAT HO G E N E S I S OF   DE L I R I U M
Many theories have been proposed to explain the etiology of
delirium, but Engel and Romano explained it best:
We thus arrive at the proposition that a derangement
in functional metabolism underlies all instances of
delirium and that this is reflected at the clinical level by
the characteristic disturbance in cognitive functions.
(Engel & Romano, 1959)
Thus we will start with the premise that delirium is a neu-
robehavioral syndrome caused by the transient disruption of
normal neuronal activity secondary to systemic disturbances
(Engel & Romano, 1959; Lipowski, 1992; Brown, 2000).
Over the years, a number of theories have been proposed in an
attempt to explain the processes leading to the development
of delirium (Maldonado, 2008a). Most of these theories are
complementary, rather than competing (see Figure 41.9). It is

Precipitants of Delirium
(e.g., infection, trauma, anesthetics, surgery, hypoxia,
hypoglycemia, metabolic derrangements)

Aging/ Melatonin/Tryptophan/
Baseline Cognition Circadian Dysregulation

Neuroinflammation Oxidative Stress Neuroendocrine

Network Neurotransmitter
Disconnectivity availability

Clinical Delirium

Neuropathogenesis of Delirium Schematics of the interrelationship of current theories on the pathophysiology of delirium and how they
Figure 41.9
may relate to each other. Each proposed theory has focused on a specific mechanism or pathologic process (e.g., dopamine excess or acetylcholine
deficiency theories), observational and experiential evidence (e.g., sleep deprivation, aging), or empirical data (e.g., specific pharmacological agents’
association with post-operative delirium; intra-operative hypoxia). Most of these theories are complementary rather than competing, with many
areas of intersection and reciprocal influence. In the end, it is unlikely that any one of these theories is fully capable of explaining the etiology or
phenomenological manifestations of delirium, but rather that their interaction leads to the biochemical derangement and, ultimately, to the com-
plex cognitive and behavioral changes characteristic of delirium. (Adapted from Maldonado, 2013).

41. D E L I R I U M • 841
likely that none of these theories by itself explains the phe-
nomena of delirium, but rather that two or more of these,
if not all, act together to lead to the biochemical derange-
ment we know as delirium. Here we will highlight four of
them. A detailed review of all has been published elsewhere
(Maldonado, 2008a, 2013).
N EU ROI N F L A M M ATORY H Y P OT H E S I S
The neuroinflammatory hypothesis suggests that acute
peripheral inflammatory processes (e.g., infections, surgery,
trauma) induce activation of brain parenchymal cells and
expression of proinflammatory cytokines and inflamma-
tory mediators in the central nervous system, which in turn
induces neuronal and synaptic dysfunction that may serve as
the substrate for the neurobehavioral and cognitive symptoms
characteristic of delirium (Godbout et  al., 2005, Dantzer,
O’Connor, Freund, Johnson, & Kelley, 2008; Cunningham
et al., 2009; Godbout & Johnson, 2009; Cerejeira, Firmino,
Vaz-Serra,  & Mukaetova-Ladinska, 2010; Cunningham,
2011)  (Figure 41.10). Previous studies have already demon-
strated the brain’s ability to monitor the presence of systemic
inflammatory processes (e.g., outside the BBB) and the devel-
opment of nonspecific physiological (e.g., fever, pain, malaise,
fatigue, and anorexia) and behavioral adaptations (e.g., anhe-
donia, lethargy, social withdrawal, depressed mood, cognitive
impairment) upon exposure to infection or inflammation col-
lectively known as “sickness behavior.”
According to the neuroinflammatory hypothesis, delir-
ium represents the CNS manifestation of a systemic disease
state that has crossed the blood–brain barrier (Maldonado,
2008b). Indeed, many of the circumstances associated with
a high occurrence of delirium (e.g., infections, postopera-
tive states) can also be associated with compromise of BBB

CNS Cytokines

Hypothalamus
CRH
AVP
TRH
GnRH

LC

LH/FSH Stressors
ACTH
TSH C2
A2
Adrenal C1
Glands
Cytokines

A1
Vagus n.

T4 T3 SNS
G luc o cor tic

Gonads
oid
s

Immune
Cells
Estrogen PNS
Thymus Spleen

Lymph
Progesterone Bone Nodes
Testosterone Marrow
DHEA
Immune System
(Cells & Organs)

Schematic Illustration of Neural-Immune Interactions Immune signaling of the CNS via systemic routes and the vagus nerve (Vagus n.)
Figure 41.10
and CNS regulation of immunity via the HPA, hypothalamic–pituitary-thyroid, and hypothalamic–pituitary–gonadal axes and the sympathetic
nervous system (SNS) and parasympathetic nervous system (PNS). Cytokine expression within the CNS is represented by asterisks within the
brain. Dotted lines represent negative regulatory pathways, and solid lines represent positive regulatory pathways. CRH: corticotrophin-releasing
hormone; AVP: arginine vasopressin; TRH: thyrotropin-releasing hormone; GnRH: gonadotropin-releasing hormone; ACTH: adrenocortico-
trophin hormone; TSH: thyroid-stimulating hormone; T4: thyroxine; T3: triio-dothyronine; LH: luteinizing hormone; FSH: follicle-stimulating
hormone; LC: locus coeruleus; Al, Cl, A2, C2: brainstem adrenergic nuclei. (Marques-Deak, Cizza, & Sternberg, 2005)

842 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
integrity. It has been postulated that several illness processes
(i.e., trauma, infections) and surgical procedures may intro-
duce triggering factors leading to the activation of the inflam-
matory cascade: extensive tissue trauma, presence of foreign
organisms or substances, elevated hormone levels, blood loss
and anemia, blood transfusions, use of extracorporeal circula-
tion, hypoxia, ischemia and reperfusion, formation of hepa-
rineprotamine complexes, and microemboli formation and
migration (reviewed by Maldonado, 2008b, 2013). In addi-
tion, other factors associated with illness and surgery may
directly lead to compromise of the BBB. For example, recent
data suggests that the use of various general anesthetics (e.g.,
sevoflurane, isoflurane) can caused marked flattening of the
surfaces of brain vascular endothelial cells along with disrup-
tion of BBB-associated tight junctions at cell margins, leading
to holes in the vascular endothelial lining and increased BBB
permeability, thus facilitating plasma influx into the brain
interstitium (Forsberg et al., 2014). The frequency and mag-
nitude of this effect increases with age, thus potentially serv-
ing as a mechanism to mediate postoperative delirium and its
increased occurrence among elderly patients (Figure 41.11).
At least four pathways through which immune signals
from the periphery are transduced to the brain have been
proposed (Figure 41.12) (Dantzer et al., 2008). In the neural
pathway, locally produced cytokines activate primary afferent
nerves (e.g., vagus). In the humoral pathway, toll-like recep-
tors on macrophage-like cells residing in the circumventric-
ular organs and the choroid plexus respond to circulating
pathogen-associated molecular patterns by producing proin-
flammatory cytokines, which enter the brain by diffusion. In
the cytokine transporters pathway, proinflammatory cyto-
kines overflowing in the systemic circulation can gain access
to the brain through these saturable transport systems at the
blood–brain barrier. The prostaglandin pathway involves
IL-1 receptors that are located on perivascular macrophages
and endothelial cells of brain venules. Activation of these IL-1
receptors by circulating cytokines results in the local produc-
tion of prostaglandin E2.
The neuroinflammatory hypothesis proposes that sev-
eral conditions associated with delirium are characterized
by activation of the inflammatory cascade with acute release
of inflammatory mediators into the bloodstream (Cerejeira
et  al., 2010). CNS resident cells react to the presence of
peripheral immune signals (cytokines and lipopolysaccharide
[LPS]), leading to production of cytokines and other media-
tors in the brain, cell proliferation, and activation of the HPA
axis (thus linking it with the neuroendocrine hypothesis, dis-
cussed next) through a complex system of interactions (Figure
41.13). These neuroinflammatory changes induce neuronal
and synaptic dysfunction (i.e., affecting various neurotrans-
mitter systems), leading to the neurobehavioral and cognitive
symptoms characteristic of delirium.
Animal studies have shown that the administration of LPS
induces sickness behavior, which requires activation of proin-
flammatory cytokine signaling in the brain (Dantzer et  al.
1998; Dantzer, 2004). Microglia are the primary recipients of
peripheral inflammatory signals that reach the brain (Figure
41.14) (Miller, Maletic,  & Raison, 2009, Miller, Haroon,
Raison,  & Felger, 2013). Activated microglia, in turn, initi-
ate an inflammatory cascade whereby release of relevant cyto-
kines, chemokines, inflammatory mediators, reactive nitrogen
species (RNS), and reactive oxygen species (ROS) induces
mutual activation of astroglia, thereby amplifying inflam-
matory signals within the CNS. Cytokines, including IL-1,
IL-6, and TNF-alpha, as well as IFN-alpha and IFN-gamma
(from T cells), induce the enzyme indoleamine 2,3 dioxy-
genase (IDO), which breaks down tryptophan (TRP), the
primary precursor of 5-HT, into quinolinic acid (QUIN), a

Proposed mechanism for link between post-surgical delirium, subsequent cognitive decline
and eventual Alzheimer’s disease

Aβ42
Blood Vessel Leak Cloud Antibody
α7nAChR
Surface
Antigen

Endocytosis Intraneuronal
Aβ42 Accumulation
Cell
Death

Dendritic
Degeneration

Short-term effects Long-term effects

Disruption of brain homeostasis Chronic Abeta42 accumulation
Abnormal synaptic activity Loss of dendrites and synapses

Exposure to anesthetics causes an immediate, short-term breakdown of BBB integrity. This leads to an influx of plasma components
Figure 41.11
into the brain tissue that causes disruption of brain homeostasis and neuronal misfiring, all culminating into the array of symptoms that hallmark
delirium. Failure to completely restore BBB integrity (most common in the elderly) may trigger long-term BBB breakdown, which can drive
chronic plasma influx, more permanent disruption of brain homeostasis, and impairment of neuronal function. Chronic binding of bloodborne
brain-reactive autoantibodies and soluble amyloid peptide (Abeta42) to neurons, and their internalization via endocytosis, are essential features of
subsequent cognitive decline and early Alzheimer’s disease pathology. (Forsberg et al., 2014)

41. D E L I R I U M • 843
 (a)

PAG

CEA PB
BNST

PVN NTS

SON
VLM
Vagal nerve

+ Cytokines
PAMPs

(b)

4th ventricle
Macrophage-
like cell

AP TLR

PAMP
BBB

Cytokines
SFO
Volume diffusion
OVLT
ME CP AP NTS

Pathways that Transduce Immune Signals from the Periphery to the Brain The brain and the immune system communicate through dif-
Figure 41.12
ferent pathways. (a) In the neural pathway, peripherally produced pathogen-associated molecular patterns (PAMPs) and cytokines activate primary
afferent nerves (e.g., vagal nerve, trigeminal nerves). (b) The humoral pathway involves circulating PAMPs that reach the brain at the level of the
choroid plexus (CP) and the circumventricular organs, where PAMPs induce the production and release of proinflammatory cytokines, likely reach-
ing the brain by diffusion. (Dantzer et al., 2008)

potent NMDA agonist and stimulator of glutamate (GLU)
release. Multiple astrocytic functions are compromised due
to excessive exposure to cytokines, QUIN, and RNS/ROS,
ultimately leading to downregulation of glutamate transport-
ers, impaired glutamate reuptake, and increased glutamate
release, as well as decreased production of neurotrophic fac-
tors. Of note, oligodendroglia are especially sensitive to the
CNS inflammatory cascade and suffer damage due to over-
exposure to cytokines such as TNF-alpha, which has a direct
toxic effect on these cells, potentially contributing to apopto-
sis and demyelination. The confluence of excessive astrocytic
glutamate release, its inadequate reuptake by astrocytes and
oligodendroglia, activation of NMDA receptors by QUIN,
increased glutamate binding and activation of extrasynaptic
NMDA receptors (accessible to glutamate released from glial
elements and associated with inhibition of brain-derived neu-
rotrophic factor [BDNF] expression), decline in neurotrophic
support, and oxidative stress ultimately disrupt neural plas-
ticity through excitotoxicity and apoptosis (Figure 41.15)
(Behan & Stone, 2000; Stone, Behan, Jones, Darlington, &
Smith, 2001; Darlington et al., 2007, 2010; Stone, Forrest, &
Darlington, 2012).
Human studies have confirmed these findings. A study of
adult patients admitted to inpatient medicine wards showed
that those who developed delirium had significantly elevated
levels (i.e., above the detection limit) of IL-6 (53% versus
31%) and IL-8 (45% versus 22%), compared with patients
who did not develop delirium, even after adjusting for infec-
tion, age, and cognitive impairment (de Rooij, van Munster,
Korevaar, & Levi, 2007). This was the first study to show a
relationship between peripherally measured cytokine levels
and delirium as a symptom of sickness behavior in acutely
admitted elderly. It also showed that cognitive function can
be impaired by a systemic infection in patients with a neu-
rodegenerative disorder such as Alzheimer’s disease, and that
this cognitive decline is preceded by raised serum levels of
IL-1h. Furthermore, aging and neurodegenerative disorders
exaggerate microglial responses following stimulation by

844 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
 Psychological Stress

Immunological Effects: Neuroendocrine Effects:

IL-1 IL-6 TNF-α CRH ACTH GnRH

Neurochemical Effects: Behavioral Effects:

NE 5-HT DA IL-1 Neuronal δ sleep appetite
Pathways sexcual activity

Liver
Acute phase Circulating
proteins NK Cells
NK activity

Adrenal Gland Adrenal Gland

Cortisol Cortisol
Macrophage /Lymphocyte
IL-1 IL-2
IL-6
Increase TNF-α

Decrease
Stimulatory Physical Stress
influence
Inhibitory
influence

Interlukins and the Stress Response Both psychological stress (e.g., stress, depression) and physical stress (e.g., infection, trauma, surgery)
Figure 41.13
can activate interleukin-1, which will trigger various phenomena both peripherally (e.g., cascade activation of other cytokines, induction of acute
phase proteins) and centrally (e.g., various immunologic, neurochemical, neuroendocrine, and behavioral effects). Feedback mechanisms occur at
several levels and include negative feedback exerted by cortisol. Abbreviations: IL = interleukin, TNF-α = tumor necrosis factor-α, CRH = corti-
cotropinreleasing hormone, GnRH = gonadotropin-releasing hormone, NE = norepinephrine, 5-HT = serotonin, DA = dopamine, NK = natural
killer. (Kronfol & Remick, 2000)

systemic immune stimuli such as peripheral inflammation
and/or infection (as described in the neural aging hypothesis).
N EU ROE N D O C R I N E H Y P OT H E S I S
The neuroendocrine hypothesis suggests excessive glucocor-
ticoid (GC) levels may induce a vulnerable state in neurons.
Glucocorticoids are hormones released during the stress
response that are well known for their immunosuppressive
and antiinflammatory properties; however, recent advances
have uncovered situations wherein they have effects in the
opposite direction (Sorrells, Caso, Munhoz,  & Sapolsky,
2009). An extensive body of work has substantiated the idea
that repeated or prolonged exposure to GCs has a deleteri-
ous impact on brain function, and has also provided evi-
dence that GCs probably contribute to age-related decline
in brain function (Goosens & Sapolsky, 2007). Others have
already demonstrated how stress (induced by illness pro-
cesses, such as hypoxia/ischemia, hypoglycemia, or seizures)
can lead to excess glucocorticoid release, which can exac-
erbate cell death by inhibiting glucose transport into cells
(Figure 41.16 (Sapolsky & Pulsinelli, 1985; Sapolsky, 1990,
1996; Dinkel, Ogle, & Sapolsky, 2002). But there is grow-
ing evidence showing that GCs may have proinflammatory
effects in the brain and can even enhance neuroinflamma-
tion at multiple levels in the pathway linking LPS exposure
to inflammation (Figure  41.17) (Munhoz, Sorrells, Caso,
Scavone,  & Sapolsky, 2010; Schiepers, Wichers,  & Maes,
2005). In addition, glucocorticoids can increase proinflam-
matory cell migration, cytokine production, and even tran-
scription factor activity in the brain (Sorrells  & Sapolsky,
2007). This can lead to a number of deleterious effects,
including inhibition of glutamate reuptake in the synaptic
cleft, inhibition of Ca+ efflux or sequestration, exacerbation
of the breakdown of cytoskeletal proteins (i.e., tau), increase

41. D E L I R I U M • 845
 Peripheral
inflammatory
response

TRP
Perivascular
macrophage ↑ Cytokines
↑ Chemokines ↑ IDO ↓ 5ΗΤ
↑ Inflammatory
mediators
↑ QUIN
↑ RNS, ROS Microglia

Astrocyte

Postsynaptic
terminal
↑ GLU
Extrasynpatic
Oligodendrocytr ↓ Trophic NMDA receptor
support ↓ GLU Reuptake activation
↑ GLU Release

TNF-alpha

↓ BDNF
Apoptosis Excitotoxicity
Demyelination Presynaptic
terminal Glutamate

Effects of the CNS Inflammatory Cascade on Neural Plasticity. Microglia are the primary recipients of peripheral inflammatory signals
Figure 41.14
that reach the brain. When activated, microglia initiate an inflammatory cascade (e.g., releasing cytokines, chemokines, inflammatory mediators,
reactive nitrogen species, and reactive oxygen species) leading to the activation of astroglia, thus amplifying inflammatory signals within the CNS.
Cytokines (e.g., IL-1, IL-6, and TNF-alpha, as well as IFN-alpha and IFN-gamma [from T cells]), induce the enzyme indoleamine 2,3 dioxygenase,
affecting tryptophan metabolism and thus increasing quinolinic acid production, a potent NMDA agonist and stimulator of glutamate release.
Eventually, there is astrocytic function compromise leading to downregulation of glutamate transporters, impaired glutamate reuptake, and
increased glutamate release, as well as decreased production of neurotrophic factors. Oligodendroglia are especially sensitive to the CNS inflam-
matory cascade which leads to apoptosis and demyelination. The confluence of these factors leads to a decline in neurotrophic support, oxidative
stress, and, ultimately, disruption of neural plasticity through excitotoxicity and apoptosis. Abbreviations: 5-HT, serotonin; BDNF, brain-derived
neurotrophic factor; CNS, central nervous system; GLU, glutamate; IDO, indolamine 2,3 dioxygenase; IFN, interferon; IL, interleukin; NMDA,
N-methyl-D-aspartate; QUIN, quinolinic acid; RNS, reactive nitrogen species; ROS, reactive oxygen species; TNF, tumor necrosis factor; TRP,
tryptophan. (Miller, Maletic, & Raison, 2009)

in reactive oxygen species, decrease in activity of antioxidant
enzymes, decreased release of inhibitory neurotransmitters
(i.e., GABA), and decreased production of neurotrophins
(i.e., BDNF).
Disturbances of the HPA system have also been found in
several studies on delirium. Studies have found that patients
experiencing postoperative delirium had an impaired
stress-regulating system, with significantly elevated mean
plasma cortisol levels compared to the preoperative baseline
and nondelirious patients (McIntosh et  al., 1985). Among
delirious patients triggered by lower respiratory tract infec-
tion, 78% were found to be nonsuppressors on the DST com-
pared to 14% of the patients without delirium (O’Keeffe &
Devlin, 1994). Among non–medically ill demented patients,
those who develop delirium exhibited significant differences
between basal cortisol levels compared to demented, non-
delirious patients (Robertsson et  al., 2001). But, the most
important finding of the study was a strong relationship
between delirium and DST pathology irrespective of age and
severity of dementia, with significant differences in post-DST
cortisol levels between patients with different degrees of
delirium:  the greater the intensity of delirium, the greater
the level of nonsuppression. Similarly, delirious patients
after stroke exhibited significantly increased activation of
the HPA system compared to those without acute confu-
sion (Olsson et  al., 1989; Olsson, Marklund, Gustafson,  &
Nasman, 1992; Fassbender et  al., 1994; Johansson et  al.,
1997; Slowik et al., 2002).
The hippocampus is a major target for these effects, with
its dense concentration of glucocorticoid receptors. The

846 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
 NH2 Indoleamine NH2
N'-formyl-
2,3-dioxygenase H NH kynurenine
HN O
O OH O OH
L-tryptophan
O
tryptophan
hydroxylase formamidase
OH
HO

NH2 NH2 N COOH

kynurenic acid
kynurenine (NMDA antagonist,
NH2 O
HN O OH O OH anticonvulsant &
5-hydroxy-
tryptohan neuroprotective)
aromatic amino
acid decarboxylase kyrnuenine mono-oxygenase (MO)
HO O NH
2
C COOH
OHC
NH2
HO2C OH NH2
serotonin OH 3-Hydroxykynurenine
2-amino-3-(3-oxoprop-
5-hydroxy- HN NH2 O 1-enyl)-fumaric acid
tryptamine kynureinase (KYNU)
(5-HT) N-acetyl
transferase non-enzymatic COOH
HO cyclization
NH2
O
H OH
3-Hydroxyanthranilic acid
N CH3
OH (toxic metabolite)
quinolinate
O OH
N-acetyl- HN N
5-HT COOH
5-hydroxyindole- O 2-Amino-3-carboxymuconic
OHC NH2
O-methyltransferase HOOC semialdehyde Picolinic acid
H3C O
Aminocarboxymuconate semialdehyde
O decarboxyiase
H (ACMSD) N COOH
N CH3 COOH
OH

HN O N COOH Quinolinate phosphoribosyl transferase (QPRT) NAD+

N
Quinolinic acid
melatonin niacin
(toxic metabolite)

Metabolic Pattern of Tryptophan, Serotonin, Melatonin, Kynurenic and Quinolinic Acid. Evidence suggest that some proinflammatory cytokines can not only induce
Figure 41.15
sickness behavior, but also enhance activity of the ubiquitous indoleamine 2, 3 dioxygenase (IDO), leading to deficient tryptophan (TRP) levels, thus a reduction in 5HT and
melatonin production, and a shift to the production of kynurenine (KYN) and other neurotoxic tryptophan-derived metabolites. (Maldonado 2013; modified from Darlington et al., 2010, and Stone,
Forrest, & Darlington, 2012)
 Emotional
Chemical Stimuli 5-HT NA
Physical – +
(+)
(–) (+) +/–
Hypothalamus Hypothalamus

CRH
CRH NTS
(+)
(–) (+)? Pituitary P
Pituitary CNS
(–) Inflammation

ACTH
–
IL-1 ACTH – + IL-1 X
(+)
Adrenal TNF-α
gland
+
(+) Adrenal cortex –
Peripheral – Cytokines
GCs
(–) Inflammation
CS +/–
IL-6

Effects on heart. Immune system

(+)
blood vessels,
gonads and other Liver
organ systems
Neuroendocrine Circuits and the Effects of Inflammation.
Figure 41.16
Various stressors can activate the HPA axis. The hypothalamus is
stimulated to secrete CRH, which leads to ACTH secretion into the
peripheral circulation. ACTH in turn triggers adrenal GC release and
production. The CRH system is inhibited by GCs in a negative feed-
back loop. TNF-κ and IL-1 are produced from inflammatory sites and
are potent activators of the HPA axis. IL-6 acts synergistically with GCs
to stimulate the hepatic secretion of acute phase proteins. Although
GCs are widely known for their anti-inflammatory actions, “(–),” more
recently also proinflammatory effects have repeatedly been reported,
“(+)?” Abbreviations: HPA = hypothalamic-pituitary-adrenal; CRH =
corticotrophin-releasing hormone; ACTH = adrenocorticotrophic hor-
mone; GCs = glucocorticoids; IL = interleukin; TNF = tumor necrosis
factor; (+)  =  enhancing; (–)  =  suppressing. (Dinkel, Ogle, & Sapolsky, 2002)
hippocampal–adrenal circuit may contribute to the ampli-
fication of deliriogenic factors. There is evidence that, rela-
tively early during the metabolic stress leading to delirium,
the hippocampus begins to malfunction. This may explain
some of the memory dysfunction and errors in informa-
tion processing, leading to the confabulations that are
commonly seen in delirious patients. The loss of normal
inhibition of adrenal steroidogenesis results in the continu-
ous secretion of peak amounts of corticosteroids, leading to
further mitochondrial dysfunction and apoptosis and fur-
ther exacerbation of the catecholamine disturbances. A key
abnormality related to cortisol excess in delirium seems to
be abnormal “shut-off” of the HPA axis tested by the DST.
In experimental models, the hippocampal formation is of
prime importance for normal HPA axis shut-off (Olsson,
1999). In this brain area, a close interaction between
neurotransmitters—notably ACH, 5HT, and NE—and
glucocorticoid receptors is relevant for the development of
delirium in elderly patients with stroke and neurodegenera-
tive brain diseases.
Inflammation, Neurotransmitters & the HPA Axis. The brain
Figure 41.17
regulates the immune system, and the immune system modulates brain
activity through various interactions. The hypothalamic–pituitary–adrenal
(HPA) axis is the main route of immunoregulation, next to (para)sympa-
thetic innervation of organs of the immune system. Immune cells secrete
cytokines, which not only regulate the immune response but also act to
block the actions of other humoral substances at different levels of the HPA
axis (black bars). (Schiepers, Wichers, & Maes, 2005)
OX I DAT I V E S T R E S S H Y P OT H E S I S
The oxidative stress hypothesis proposes that decreased
brain oxidative metabolism leads to abnormalities of vari-
ous neurotransmitter systems, causing cerebral dysfunction
and the behavioral symptoms of delirium. Thus, severe ill-
ness processes, combined with both decreased oxygen sup-
ply and/or increased oxygen demand, may lead to the same
common end problem; namely, decreased oxygen availabil-
ity to cerebral tissue. This has been the subject of an exten-
sive review described elsewhere Figure 41.18 (Maldonado,
2013). In summary, hypoxia triggers diverse reconfigura-
tions of widespread neuronal network at all levels of the
nervous system (i.e., molecular, cellular, synaptic, neuronal,
network): synaptic transmission is depressed through pre-
synaptic mechanisms and excitatory/inhibitory alterations
involving K+, Na+, and Ca 2+ channels. Cerebral ischemia
leads to a rapid depletion of energy stores, triggering a com-
plex cascade of cellular events, including cellular depolar-
ization and Ca2+ influx, resulting in excitotoxic cell death
(see Figure 41.19).
Inadequate oxidative metabolism may be one of the causes
of the problems observed in delirium; namely, inability to
maintain ionic gradients causing “spreading depression”;
abnormal neurotransmitter synthesis, metabolism and release;
and a failure to effectively eliminate neurotoxic byprod-
ucts. Decreased oxygenation causes a failure in oxidative

848 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
 A Basic Pathoetiological Model of Delirium. This figure illustrates the theorized intersections between the oxidative stress hypothesis (OSH) and the neurotransmitter hypothesis (NTH) of the
Figure 41.18
delirium, demonstrating potential common biochemical outcomes, which may explain the complex cognitive and behavioral changes characteristic of delirium. The OSH proposes that a number of physiolog-
ical processes (e.g., hypoxia, severe illness, infectious processes) may give rise to increased oxygen consumption and/or decreased oxygen availability, with associated increased energy expenditure and reduced
cerebral oxidative metabolism, leading to cerebral dysfunction and associated cognitive and behavioral symptoms of delirium. The NTH was proposed after clinical observations that delirium occurred after
the use of substances (e.g., medications, toxins) that alter neurotransmitter function or availability. The OSH intersects with the NTH as decreased oxygenation causes a failure in oxidative metabolism, lead-
ing to a failure of the ATP-ase pump system, which leads to an inability to maintain adequate ionic gradients, which in turn leads to significant electrolyte alterations (e.g., influx of Na+ and Ca2+; efflux of
K+) and subsequent alterations (e.g., excess release or decreased availability) of several neurotransmitters (e.g., GLU, DA, Ach). (Maldonado 2008b, 2013)
 Decreased cerebral perfusion
Depletion in energy stores

Cerebral Ischemia

Apoptosis Excitotoxicity Inflammation (Programmed Cell Death)

Cell shrinkage Cytokine (IL-1, TNF-α) production
Chromatin aggregation Chemokines
Preservation of cell membrane EAAs (Glutamate, Aspartate), Adhesion molecules (selections, integrins, ICAM-1)
integrity Catecholamine Proteases
Preservation of mitochondria BBB breakdown
Lack of inflammation
Activation of ionophore-linked channels Activation of Metabotropic
(NMDA, AMPA) receptors

Influx of intracellular Ca2+

Protein phosphorylation Modification

Proteolysis
NO, oxygen free radical production
Altered gene expression

Neuronal Cell Death

Schematic diagram representing events leading to ischemic brain injury.
Harukuni & Bhardwaj, Neurol Clin 2006.

Mechanisms of Brain Injury after Global Cerebral Ischemia. During cerebral ischemia, excess glutamate exits into the extracellular com-
Figure 41.19
partment due to cellular depolarization, coupled with its impaired uptake, which results in increases in intracellular Ca2+. The cascade of events
responsible for glutamate excitotoxicity includes three distinct processes: (1) induction, whereby extracellular glutamate efflux is transduced by
receptors on the neuronal membrane to cause intracellular Ca2+ overload, which leads to lethal intracellular derangements; (2) amplification of the
derangement, with an increase in intensity and involvement of other neurons; and (3) expression of cell death triggered by cytotoxic cascades. Excess
release of Ca2+ and its intracellular influx is thought to be the primary trigger for a variety of complex, deleterious intracellular processes that result
from activation of catabolic enzymes such as phospholipases (which lead to cell membrane breakdown, arachidonic acid, and free radical formation)
and endonucleases (which lead to fragmentation of genomic DNA and energy failure due to mitochondrial dysfunction). (Harukuni & Bhardwaj, 2006)

metabolism, which leads to a failure of the Adenosine triphos-
phate (ATP)-ase pump system. When the pump fails, the ionic
gradients cannot be maintained, leading to significant influxes
of sodium (Na+) followed by calcium (Ca2+), while potas-
sium (K+) moves out of the cell. Some have theorized that
it is the excess inward flux of Ca2+ that precipitates the most
significant neurobehavioral disturbances observed in deliri-
ous patients. The influx of Ca2+ during hypoxic conditions
is associated with the dramatic release of several neurotrans-
mitters, particularly glutamate (GLU) and dopamine (DA).
Glutamate further potentiates its own release as GLU stimu-
lates the influx of Ca2+, and it accumulates in the extracellular
space as its reuptake and metabolism in glial cells is impeded
by the ATP-ase pump failure. In addition, at least two factors
facilitate dramatic increases in DA:  first, the conversion of
DA to norepinephrine (NE), which is oxygen dependent, is
significantly decreased; second, the catechol-o-methyl trans-
ferase (COMT) enzymes, required for degradation of DA,
get inhibited by toxic metabolites under hypoxic conditions,
leading to even greater accumulation of DA (Graham, 1984).
At the same time, serotonin (5HT) levels fall moderately in
the cortex, increase in the striatum, and remain stable in the
brainstem (Broderick & Gibson, 1989).
Hypoxia also leads to a reduced synthesis and release of
acetylcholine, especially in the basal forebrain cholinergic
centers. Indeed, cholinergic neurotransmission is particularly
sensitive to metabolic insults, such as diminished availability
of glucose and oxygen. ACh synthesis requires acetyl coen-
zyme A, which is a key intermediate linking the glycolytic
pathway and the citric acid cycle. Thus, reduction in cerebral
oxygen and glucose supply and deficiencies in enzyme cofac-
tors such as thiamine may induce delirium by impairing ACh
production.
There are definite data correlating poor oxygenation and
cerebral dysfunction. Studies have demonstrated a strong
correlation between intraoperative oxygen saturation and
postoperative mental function (Rosenberg & Kehlet, 1993).
Studies have demonstrated that delirium can be induced
in healthy control subjects by dropping PaO2 to 35 mmHg
(Gibson  & Peterson, 1981). A  study of ICU patients dem-
onstrated that three measures of oxygenation (i.e., hemo-
globin level, hematocrit, pulse oximetry; see Figure 41.4)
were worse in the patients who later developed delirium,
and that clinical factors associated with greater oxidative
stress (e.g., sepsis, pneumonia) occurred more frequently
among those diagnosed with delirium (Seaman et al., 2006).

850 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
A  recent study demonstrated that intraoperative cerebral
oxygen desaturation was a significant risk factor for post-
operative delirium in CABG patients (Figure 41.5) (Slater
et al., 2009).
DI U R N A L DY S R E GU L AT ION OR M E L ATON I N
DY S R E GU L AT ION H Y P OT H E S I S
The diurnal dysregulation or melatonin dysregulation
hypothesis suggests that the 24-hour internal “clock” (cir-
cadian pattern) is maintained by environmental factors,
primarily light exposure, which affect melatonin secretion
(BaHammam, 2006). Sleep deprivation may lead to the devel-
opment of memory deficits (Walker  & Stickgold, 2004). In
fact, “chronic partial sleep deprivation” (i.e., sleeping limited
to 4 hours per night, for 5 consecutive nights) has been shown
to lead into cumulative impairment in attention, critical
thinking, reaction time, and recall (Dinges, 2006). In turn,
sleep deprivation (even just 36 consecutive hours) may lead to
symptoms of emotional imbalance (i.e., short temper, mood
swings, and excessive emotional response). Cumulative sleep
debt can cause delirium in itself, and can aggravate or perpet-
uate delirium (Ito, Harada, Hayashida, Ishino, & Nakayama,
2006; Pandharipande  & Ely, 2006). Similarly, others have
found sleep deprivation to consistently precede onset of delir-
ium in cardiac surgical patients (Sveinsson, 1975), and that
ICU patients with sleep deprivation were significantly more
likely to develop delirium than patients without sleep depriva-
tion (Helton, Gordon, & Nunnery, 1980).
Melatonin plays important roles in multiple bodily func-
tions (Reiter, 1991; Brzezinski, 1997; Verster, 2009). It has a
chronobiotic effect (affecting aspects of biological time struc-
ture), sleep–wake cycle regulatory effects, and helps reset
circadian rhythm disturbances. It has extensive antioxidant
activity (with a particular role in the protection of nuclear
and mitochondrial DNA), extensive anti-inflammatory activ-
ity, and some anti-nociceptive and analgesic effects; and data
suggest that melatonin receptors appear to be important in
mechanisms of learning and memory. Melatonin also inhibits
the aggregation of the amyloid beta protein into neurotoxic
micro-aggregates responsible for the neurofibrillary tangles
characteristic of Alzheimer’s disease, and it prevents the
hyperphosphorylation of the tau protein. All of these factors
may have potential implications regarding the development of
delirium in the medically ill and postoperative patient.
In addition, current evidence suggests that acute and
chronic sleep deprivation is associated with decreased propor-
tions of natural killer cells, lower antibody titers following
influenza-virus immunization, reduced lymphokine-activated
killer activity, and reduced IL-2 production. Conversely, cyto-
kines may play a role in normal sleep regulation by increasing
non-REM sleep and decreasing REM sleep, and, during inflam-
matory events, an increase in cytokine levels may intensify their
effects on sleep regulation. Moreover, sleep deprivation may
alter endocrine and metabolic functions, altering the normal
pattern of cortisol release and contributing to impairment of
the regulation of glucocorticoid secretion via negative feedback
mechanisms, insulin resistance, and glucose intolerance.
41. D E L I R I U M
Several studies have demonstrated a relationship between
melatonin secretion and delirium. A study of elderly patients
after major abdominal surgery had their plasma melatonin
levels measured preoperatively and every 2 hours during
the postoperative period, and demonstrated that those who
did not develop delirium did not have significant changes
in plasma melatonin levels compared with preoperative val-
ues (Shigeta, Yasui, et al., 2001). Furthermore, patients who
experienced postoperative delirium could be broken down
into two groups:  those who became delirious postopera-
tively, with an otherwise uncomplicated recovery; and those
who developed postoperative complications such as pneumo-
nia. Patients who went on to become delirious, but without
other complications, demonstrated reduced plasma melato-
nin levels. Those delirious but with other complications had
elevated plasma melatonin levels, similar to that seen in early
sepsis (Mundigler et al., 2002). Similarly, a study of subjects
undergoing thoracic esophagectomy demonstrated a signifi-
cant correlation between ICU psychosis (i.e., delirium) and
an irregular melatonin circadian rhythm (i.e., abnormally low
serum levels of melatonin), measured every 6 hours during the
first 4 postoperative days (P =.0001) (Miyazaki et al., 2003).
Of note, some have observed a relationship between the
motoric delirium subtype and melatonin levels. A  study of
hospitalized elderly medical patients evaluated them daily
using the Delerium Rating Scale (DRS) and conducted uri-
nary measures of their 6-sulphatoxymelatonin (6-SMT), the
chief metabolite of melatonin and a reliable proxy for serum
melatonin (Matthews, Guerin,  & Wang, 1991). The mean
and standard deviation values of 6-SMT for each of the three
motoric subtypes were calculated, and it was found that dur-
ing periods of hyperactive delirium, subjects had decreased
urinary 6-SMT levels compared with when they had recov-
ered. Patients with hypoactive delirium had raised 6-SMT
levels compared with when they had recovered, while those
with the mixed form of delirium had no difference in urinary
metabolite concentrations (Balan et al., 2003).
N EU RON A L AG I NG H Y P OT H E S I S
Finally, the neuronal aging hypothesis suggests that elderly
patients are at increased risk for developing delirium, based
on the fact that aging is associated with age-related cerebral
changes in stress-regulating neurotransmitter and intracel-
lular signal transduction systems, and observations suggest-
ing that chronic neurodegeneration is accompanied by an
inflammatory response characterized by a chronic, but selec-
tive, activation of CNS microglial cells that are “primed” to
produce exaggerated inflammatory responses to immuno-
logical challenges (Figure 41.20) (Cunningham et al., 2005,
2009). This inflammation is assumed to contribute to disease
progression through the production of inflammatory medi-
ators. Data suggest that the aging process is associated with
a twofold to fourfold increase in baseline levels of circulat-
ing inflammatory mediators, including cytokines and acute
phase proteins (Rosczyk, Sparkman,  & Johnson, 2008).
This may help explain why even minor surgical trauma was
associated with increased IL-1b hippocampal expression in
• 851
 (a) Periphery: acute insult Vulnerable brain: aging/dementia

Systemic Endothelial cells of

Cerebral vasculature Primed
Inflammatory Tissue Macrophages Microglia
Event
Infection, Injury
Surgery IL-1β
TNF-α
PAMPS INDIRECT IL-1β
IL-6
IL-1β TNF-α
IFN α/β
PGE2 PGE2
PGE2
TNF-α NO
IFN α/β DIRECT
GCs

Neuronal Neuronal
(b) Systemic inflammatory events Neuronal
Dysfunction Damage Death
IL-1RI
70
TNFp55R
60
GR EP1-4
50
Functional decline

Reversibly impaired
40 neuronal function Reversible Irreversible
30

20 disease+SIEs
disease w/o SIEs
10
normal SIEs
LONG-TERM
0 DELIRIUM COGNITIVE
12

13

14

y
15

16

y
17

18

y
19

DECLINE
er

er

er
SI

SI

SI
ov

ov

ov
c

c
re

re

re

Disease progression (weeks)

Effect of Inflammation on the Aging Brain. (a) Systemic inflammatory events trigger the release of inflammatory mediators by tissue mac-
Figure 41.20
rophages and brain vascular endothelial cells. These mediators may affect neuronal function directly, or via the activation of microglial cells that have
become primed by neurodegenerative disease or aging. Inflammatory mediators may cause reversible disruption of neuronal function, leading to various
neurobehavioral syndromes including delirium. These mediators may also induce acute neuronal synaptic or dendritic damage that may be reversible
and contribute to delirium, or may be irreversible and contribute to long-term cognitive decline. (b) Successive systemic inflammatory insults induce
acute dysfunction, which is progressively less reversible each time, and also contribute to the progression of permanent disability. Abbreviations: IL-1RI,
interleukin 1 receptor type I; TNFp55, TNFp55 receptor; GCs, glucocorticoids; GR, glucocorticoid receptor; NO, nitric oxide; EP1–4, prostaglandin
receptors 1–4; PAMPs: pathogen-associated molecular patterns; IFNα/β, interferon α/β; SIEs, systemic inflammatory events.

animal studies. Aging is also associated with decreased vol- published meta-analysis on the matter found a positive asso-
ume of ACh-producing cells and decreased cerebral oxida- ciation between POCD and the APOE sigma4 allele.
tive metabolism. Thus, the decline in cognitive functioning In the end, any or all of the above theories lead to the same
associated with the normal aging process is aggravated by common result: changes in neurotransmitter concentration or
the presence of even mild hypoxia, which further inhibits receptor sensitivity (due to any of the mechanisms discussed
ACh synthesis and its release. As previously discussed, the by the previous hypotheses) mediate the different symptoms
probability of experiencing delirium rises with age. Of note, and clinical presentations of delirium. In general, the most
some have suggested that prostaglandins are responsible for commonly described neurotransmitter changes associated
the LPS-induced inhibition of sickness behaviors, either with delirium are excess release of norepinephrine (NE),
independent of cytokine activity or as complement (i.e., sen- dopamine (DA), and/or glutamate (GLU), and increased
sitizing certain targets to the effect of cytokines) (Teeling Ca+ channel activity (Ca+ Ch); reduced availability of ace-
et al., 2007). tylcholine ( Ach) and/or melatonin (MEL); and either
There has been some controversy about the contribution of a decreased and increased activity in serotonin (5HT),
apolipoprotein E and the development of postoperative cogni- histamine (H1&2), and/or gamma-amino butyric acid
tive dysfunction (POCD). To date, there have been nine pub- (GABA), probably depending on the etiology or motoric
lished papers on the topic; four of them found that APOE-E4 presentation (see Table 41.7).
genotype was not associated with POCD (Abildstrom et al.,
2004; Tagarakis et al., 2007; van Munster, Korevaar, de Rooij,
Levi,  & Zwinderman, 2007; Bryson et  al., 2011), while the C L I N IC A L PR E S E N TAT ION OF   DE L I R I U M
other five suggested a positive relationship between POCD
and APOE genotype (Adamis et  al., 2007; Ely et  al., 2007; Delirium is an acute or subacute organic mental syndrome
Leung et al., 2007; Zhang et al., 2008; van Munster, Korevaar, characterized by disturbance in attention (often evidenced
Zwinderman, Leeflang, & de Rooij, 2009). (For a summary by a reduced ability to direct, focus, sustain, and shift atten-
of all published studies see Table 41.6.) Of note, the only tion) and awareness (with impaired orientation to the

852 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
Table 41.6 ASSOCIATION BET W EEN APOLIPOPROTEIN E GENOTYPE AND POSTOPER ATIVE COGNITIVE
DYSFUNCTION
STUDY(N = 10) POPULATION
Abildstrom et al., 2004 Patients aged 40 yrs.
Prospective, n = 967 and older undergoing
non-cardiac surgery
Adamis et al., 2007 ≥70 y/o hospitalized
Prospective, n = 164 medically ill patients
Leung et al., 2007 nested cohort study;
Prospective, n = 190 patients aged ≥ 65 yr,
scheduled to undergo
major noncardiac surgery
requiring anesthesia.
Tagarakis et al., 2007, Elderly adults undergoing
n = 137 cardiac bypass surgery;
excluded dementia
Van Munster et al., 2007 Acutely admitted patients
n = 415 to the Department of
Medicine of 65 yrs
and over
Ely et al., 2007 Mechanically ventilated
n = 53 ICU patients
Zhang et al., 2008 Elderly patients (>60 yo)
n = 196 scheduled for major
abdominal surgery
requiring general
anesthesia
Van Munster et al., 2009 Medical department and
n = 656 + orthopedic/traumatology
Meta-analysis department of university
hospital from 2003 to 2007.
Bryson et al., 2011 Patients ≥60 yrs. of age
n = 88 undergoing open aortic
repair.
FINDINGS P-VALUE
• One week after surgery, the incidence of POCD was 11.7% P = 0.41
in patients with the epsilon 4 allele and 9.9% in patients
without the epsilon4 allele.
• Conclusion: unable to show a significant association
between apolipoprotein E genotype and POCD.
• Recovery was significantly (P <.05) associated with lack of p < 0.05
APOE 4 allele and higher initial interferon (IFN)-gamma.
• A model incorporating gender, APOE epsilon4 status and
IGF-I levels predicted recovery or not from delirium in
76.5% of cases, with a sensitivity 0.77 and specificity 0.75.
• It further found a positive relationship between delirium
with APOE genotype, IFN-gamma, and IGF-I, but not
with IL-6, IL-1, TNF-alpha, and leukemia inhibitory
factor.
• The presence of one copy of the e4 allele was associated P = 0.005
with an increased risk of early postoperative delirium
(28.3% vs. 11.1%; P = 0.005). Even after adjusting for
covariates, patients with one copy of the e4 allele were still
more likely to have an increased risk of early postoperative
delirium (OR, 3.64; 95% CI, 1.51–8.77) compared with
those without the e4 allele.
• Study confirmed the high incidence of cognitive decline NS
and delirium after coronary surgery, but it does not
support the role of the APOE-epsilon 4 allele in the
occurrence of delirium.
The OR for carriers of an APOE epsilon 4-allele compared P = 0.12
with patients without an APOE epsilon 4-allele for
developing delirium was 1.17 (95% CI: 0.49–2.78) in the
cognitively intact patients and 0.42 (95% CI: 0.14–1.30)
in the cognitively impaired patients. No relation existed
between the total number of APOE epsilon 4-alleles and the
different delirium subtypes (P = 0.12).
Using multivariable regression analysis to adjust for age, p = .005
admission diagnosis of sepsis or acute respiratory distress
syndrome or pneumonia, severity of illness, and duration
of coma, the presence of APOE4 allele was the strongest
predictor of delirium duration (OR, 7.32; 95% CI,
1.82–29.51, p =.005).
The presence of the e4 allele and low level of education were P = 0.01
both associated with an increased risk of EA (36.9% vs.
15.8%, P = 0.005; 30% vs. 14.3%, P = 0.01). After adjustment
for covariates, the patients with the copy of the e4 allele were
shown to have a greater likeliness of an increased risk of EA
(OR: 4.32; 95% CI: 1.75–10.05).
The OR for delirium adjusted for age, cognitive, and P = 0.04
functional impairment of sigma4 carriers compared with
non-sigma4 carriers was 1.7 (95% CI: 1.1–2.6). Four studies
were added to the meta-analysis, which included 1,099
patients in total. The OR for delirium in the meta-analysis
was 1.6 (95% CI: 0.9–2.7) of sigma4 carriers compared with
non-sigma4 carriers.
This study and meta-analysis suggest an association between
delirium and the APOE sigma4 allele.
Delirium predicted POCD at discharge (OR 2.86; 95% CIs p = 0.625
0.99 to 8.27) but not at three months. Apolipoprotein E-ε4
genotype was not associated with either delirium or POCD
following adjustment for covariates.
Table 41.7 THEOR IZED NEUROCHEMICAL MECHANISMS ASSOCIATED WITH CONDITIONS LEADING TO DELIR IUM

HPA NMDA CHANGES

DELIR IUMSOURCE ACH DA GLU GABA 5HT NE TR P PHE HIS CYTOK A XIS ACTIVITY IN R BF EEG MEL INFLA M CORT

Anoxia/hypoxia         , ╬ ╬  ╬    

Aging          ╬ ╬  ╬    

TBI          ╬      ╬ 

CVA          ╬   ╬   ╬ 

Hepatic failure          ╬ ╬  ╬    
(encephalopathy)

Sleep deprivation   ╬        ╬    ╬ ╬ 

Trauma, Sx,             ╬    
& Post-op

ETOH and           ╬      
CNSdepressants
withdrawal

Infection/sepsis           ╬ ╬ ╬    

Dehydration and       ? ?   ╬   ╬  ╬ 
electrolyte imbalance
Medical illness    ╬         ╬ ╬  ╬ 
Legend & Abbreviations:  = likely to be increased or activated;  = likely to be decreased or slowed;  = no significant changes; (╬) = probably a contributor, exact mechanism is unclear; (−) = likely not 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-amino butyric acid; 5HT = 5-hydroxytryptamine or serotonin; NE = norepinephrine; Trp = tryptophan; Phe = phenylalanine; His = histamine; Cytok = cytokines; HPA axis = hypot
halamic-pituitary-adrenocortical axis; NMDA = N-methyl-D-aspartic acid; RBF = regional blood flow; EEG = electroencephalograph; Mel = melatonin; Inflam = inflammation; Cort = cortisol.
Source: Adapted from Maldonado, J. R. (2013). Neuropathogenesis of delirium: Review of current etiologic theories and common pathways. American Journal of Geriatric Psychiatry, 21, 1190–1222.
environment; Criterion A; APA, 1994). An additional dis-
turbance is present in cognition (e.g., memory deficit, disori-
entation), language, visuospatial ability, or perception (e.g.,
hallucinations or delusions; Criterion C; APA, 1994). Often,
the syndrome is accompanied by a disordered sleep–wake
cycle, (e.g., sleep–wake cycle reversal, sundowning) and alter-
ations in psychomotor activity (e.g., decreased or increased
depending on the type of delirium; APA, 1994). By defini-
tion, these changes develop over a relatively short period of
time (i.e., hours to days) and represent a change from baseline
attention and awareness, and have a tendency to fluctuate in
severity during the course of a day (e.g., waxing and waning;
Criterion B; APA, 1994). Similarly, the presentation cannot
be better explained by a preexisting psychiatric disorder (e.g.,
schizophrenia or dementia) and are deemed to be caused as
a result of a primary medical disorder and/or its treatment,
or mediated by psychoactive substances (e.g., acute use or
intoxication, or substance withdrawal, as in delirium tremens;
Criteria D & E; APA, 1994).
Some have suggested there are three core domains of
delirium: cognitive deficits (characterized by attention and vig-
ilance disturbances); circadian rhythm dysregulation (charac-
terized by fragmentation of the sleep–wake cycle); and higher
cortical dysfunction (characterized by semantic and compre-
hension deficits, executive dysfunction, and thought process
disturbances) (Meagher et al., 2007). In fact, sleep–wake cycle
disturbance, inattention, and language/thought process/com-
prehension abnormalities were the most frequent, consistent,
and differentiating symptoms of delirium in medical/surgi-
cal populations (Trzepacz et al., 2001; Meagher et al., 2007;
Franco, Trzepacz, Mejia, & Ochoa, 2009).
The clinical features of delirium include a prodromal
phase, usually marked by restlessness, anxiety, irritability;
and the experiencing of nightmares, transient hallucina-
tions (often hypnagogic), and sleep disturbances, which usu-
ally develop over a period of several hours to days before the
patient exhibits the full-blown syndrome. Often patients
recover from a comatose or medication-induced state (e.g.,
midazolam or propofol infusion) in a confused and agitated
state (i.e., emergence delirium).
Even though delirium was originally defined as a transient
syndrome, “chronic” forms may be seen in patients experienc-
ing protracted medical problems (e.g., chronic infections),
those with concomitant psychiatric disorders (e.g., bipolar
disease, psychosis), those with baseline cognitive impairment
(e.g., delirium superimposed on dementia, traumatic brain
injury [TBI]), and patients with combined delirium and
CNS-depressant withdrawal (e.g., delirium tremens, opioid
or benzodiazepine withdrawal) (see “Impact of Delirium”
section). More commonly, patients exhibit a rapid and fluctu-
ating course with symptoms varying rapidly over time, giving
rise to the classic waxing and waning picture. Patients often
experience the appearance or exacerbation of behavioral dis-
turbances in the evening and at night, leading to the classic
“sundowning” syndrome (Cardinali, Brusco, Liberczuk,  &
Furio, 2002). Whether this represents a function of biologi-
cal circadian rhythm or results from decreased sensory input
is unclear. The appearance of lucid intervals in the clinical
41. D E L I R I U M
course is an important observation and may even aid in the
diagnosis of delirium.
As described above, global disturbance of cognition is
a cardinal feature of delirium. All “main cognitive func-
tions” (e.g., perception, thinking, memory) are impaired or
abnormal to some extent. In addition, delirious patients have
difficulty sustaining attention and responding to stimuli
selectively. Usually, patients are easily distracted by their sur-
rounding environment. Delirium may be better understood
as a disorder of awareness, not arousal. Yet, arousal problems
may be associated with the various presentations of delirium.
For example, in the case of hypoactive delirium, the activating
system may be hypoactive, in which case the patient would
appear apathetic, somnolent, and quietly confused. In other
patients, the brainstem’s activating system may be hyperac-
tive, in which case the patient is agitated and hypervigilant,
exhibiting psychomotor hyperactivity and agitation, leading
to the hyperactive type of delirium.
Often, the sleep–wake cycle is disturbed. This is not only
symptomatic of delirium, but in itself may either cause or
exacerbate the condition due to sleep deprivation. More com-
monly, patients exhibit sleep–awake cycle reversal: night sleep
is usually shortened and fragmented; while daytime wakeful-
ness is often reduced, presenting in the form of drowsiness
and the tendency to nap or being difficult to arouse during
the day.
Problems with orientation are likely to be associated
with the patients’ inability to adequately process incom-
ing information and their impaired memory formation. In
fact, delirious patients seem to experience a reduced ability
to discriminate and integrate perceptions. This may include
a reduced ability to relate incoming stimuli meaningfully to
previously acquired knowledge. Thus, often patients have a
difficult time telling apart perceptions, images, dreams, illu-
sions, and hallucinations. This may be the reason why patients
commonly report their experiences as happening “as if dream-
ing while still awake.” Often patients experience emotional
dysregulation or lability.
Delirious patients’ memory is often impaired in most
aspects (i.e., registration, retention, and recall). They may
exhibit various degrees of both anterograde and retrograde
amnesia, though their remote memory is often spared.
Delirious patients often exhibit a disorganized and frag-
mented thinking pattern. The disturbances range from an
inability to direct thoughts at will, to a complete inability to
think coherently. This includes an inability to reason, solve
problems, anticipate the consequences of actions, and grasp
the meaning of abstract words. This disorganization of think-
ing, coupled with memory disturbance and misinterpreta-
tion of environmental stimuli, often leads to the development
of delusional thinking (usually of paranoid nature) and
confabulation.
DE L I R I U M S U B T Y PE S
Historically, the first delirium subtypes have been based on
behavioral and motoric characteristics (Liptzin  & Levkoff,
1992). Factor analysis has confirmed the existence of at least
• 855
two different clusters of symptoms:  hyperalert/hyperactive
features (e.g., agitation, hyperreactivity, aggressiveness, hallu-
cinations, delusions); and the hypoalert/hypoactive features
(e.g., decreased reactivity, motor and speech retardation, facial
inexpressiveness) (Camus et al., 2000).
Subsequent studies have suggested that there are at least
three types of delirium, based on their clinical (motoric) man-
ifestations:  hyperactive, hypoactive, and mixed (Meagher,
O’Hanlon, O'Mahony, Casey, & Trzepacz, 2000; Meagher &
Trzepacz, 2000). It is important to recognize that delirium
seems to manifest differently in different populations under
study (see Figure 41.21). For example, studies in the general
adult medical population revealed that, among referrals to
a consultation psychiatry service, the most common type
appeared to the mixed type (46%), followed by the hyper-
active (30%) and the hypoactive (24%) among those meet-
ing International Classification of Diseases–10 (ICD-10)
(Box 41.3) criteria for delirium (Meagher et al., 2000). Yet, the
numbers seem quite different when we look at the critically
ill populations. Among medical ICU (MICU) patients, the
most common type continues to be the mixed type (54.9%),
but there is a significant reversal on the remaining types, with
a much larger percentage of hypoactive (43.5%) than hyperac-
tive (1.6%) types (Peterson et al., 2006). Meanwhile, among
surgical ICU (SICU) patients, there is a significant reversal
on the types of delirium found, with the most common form
of delirium being the hypoactive type (64%), followed by the
mixed type (27%), and only a minority of subjects exhibiting
the purely hyperactive (9%) type (Pandharipande, Cotton
et al., 2007). Finally, a study of elderly patients in the medicine
wards found that the significant majority of subjects experi-
enced the hypoactive subtype (65%), compared to the hyper-
active (25%) or mixed (10%) subtypes (Khurana et al., 2011).
What is most significant about these findings is not
only that hypoactive delirium is often missed or misdiag-
nosed (Liptzin  & Levkoff, 1992), but that these patients
Hyperactive (3 or more)2
• Hypervigilance
• Restlessness
• Fast/loud speech
• Anger/irritability
• Combativeness
• Impatience
• Uncooperative
• Laughing
• Swearing/singing
• Euphoria
• Wandering
• Easy startling
• Distractibility Hypoactive
• Nightmares
• Persistent thoughts
2%
54%
44%
2% MICU3
Figure 41.21 Motoric subtype of delirium depending on population under study.
856 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
experienced the worst prognosis (e.g., prolonged hospi-
tal stay, higher morbidity and mortality) (Kiely, Jones,
Bergmann, & Marcantonio, 2007). Frequently these patients
present with depressive-like symptoms (e.g., unawareness of
the environment, lethargy, apathy, decreased level of alert-
ness, psychomotor retardation, decreased speech produc-
tion, and episodes of unresponsiveness or staring). Patients
with hypoactive delirium often endorsed depressive symp-
toms, such as low mood (60%), worthlessness (68%), and
frequent thoughts of death (52%) (Farrell  & Ganzini,
1995). Studies have demonstrated that a large percentage of
these patients are inappropriately diagnosed and treated as
depressed. Hypoactive delirium is unrecognized (and there-
fore untreated or mismanaged) in 66%–84% of patients. In
fact, studies have found that patients with hypoactive delir-
ium demonstrated a seven-fold risk of under-recognition by
nurses and that older, visually impaired, demented patients
with hypoactive delirium were 20 times less likely to be rec-
ognized by nurses as having delirium (Inouye, Foreman,
Mion, Katz,  & Cooney, 2001). Our own experience at
Stanford University Hospital, in parallel to that of others
(Farrell  & Ganzini, 1995; Kishi et  al., 2007), found that
42% of the time when the psychiatry consultation service
was called to treat a patient for “depression,” the patient’s
correct diagnosis was hypoactive delirium (Maldonado,
Dhami,  & Wise, 2003). The same study found that nearly
80% of these patients had been inappropriately prescribed
antidepressant medications.
To most physicians, the clearest and most recognizable
form is the hyperactive type. Most clinicians agree that a
confused, disoriented patient who does not have a preexist-
ing psychiatric diagnosis, who suddenly becomes agitated,
combative, and/or assaultive, is probably suffering from the
hyperactive or “agitated type” of delirium. We use the term
“mixed type” to describe the classic “waxing and waning”
pattern, commonly seen in medically ill patients who appear
Delirium Subtypes
Hypoactive (4 or more)2
General Poputation1
☐ Unawareness
☐ Lethargy
46% ☐ Decreased alertness
☐ Staring
Hyperactive24%
☐ Sparse/slow speech
30% ☐ Apathy
☐ Decreased motor
activity
Hyperactive Mixed
34% 25% 10%
64% 65%
SICU4 Hospitalized Elderly5
 Box 41.3 ICD-10 DIAGNOSTIC CRITERIA FOR DELIRIUM

For a definite diagnosis, delirium symptoms, mild or severe, should be present in each one of the following areas:
1. Impairment of consciousness and attention, on a continuum from clouding to coma (e.g., reduced ability to direct, focus, sustain, and
shift attention);
2. Global disturbance of cognition (e.g., perceptual distortions, illusions and hallucinations—most often visual; impairment of abstract
thinking and comprehension, with or without transient delusions, but typically with some degree of incoherence; impairment of
immediate recall and of recent memory but with relatively intact remote memory; disorientation for time as well as, in more severe
cases, for place and person);
3. Psychomotor disturbances (e.g., hypo- or hyperactivity and unpredictable shifts from one to the other; increased reaction time;
increased or decreased flow of speech; enhanced startle reaction);
4. Disturbance of the sleep-wake cycle (e.g., insomnia or, in severe cases, total sleep loss or reversal of the sleep-wake cycle; daytime
drowsiness; nocturnal worsening of symptoms; disturbing dreams or nightmares, which may continue as hallucinations after
awakening);
5. Emotional disturbance (e.g. depression, anxiety or fear, irritability, euphoria, apathy, or wondering perplexity).
The onset is usually rapid, the course diurnally fluctuating, and the total duration of the condition less than six months. The above
clinical picture is so characteristic that a fairly confident diagnosis of delirium can be made even if the underlying cause is not clearly
established. In addition to a history of an underlying physical or brain disease, evidence of cerebral dysfunction (e.g., an abnormal elec-
troencephalogram, usually but not invariably showing a slowing of the background activity) may be required if the diagnosis is in doubt.

agitated and combative at times, with alternating episodes of
somnolence and hypoactivity.
In addition to motoric subtypes, researchers have recently
described a new delirium presentation form:  subsyndromal
delirium (SSD). Different from the other three described
above, this is not differentiated by motor type but by an
incomplete presentation of diagnostic criteria, along with
cognitive impairment. In one study, patients with SSD
(defined as meeting ≥ 2 DSM-delirium criteria, but did not
meet Confusion Assessment Method [CAM] diagnostic
criteria for delirium) had outcomes similar to or worse than
those with mild CAM-defined delirium (e.g., nursing home
placement or death at 6 months: 27% vs. 0%) (Marcantonio,
Ta, Duthie,  & Resnick, 2002). Others have demonstrated
that when the appropriate screening or diagnostic test is used
to allow differentiation between no delirium, subsyndromal
delirium, and clinical delirium (e.g., severity scale), patients
with SSD in the general medical wards experienced longer
acute care hospital stays, increased post-discharge mortal-
ity, more symptoms of delirium, and a lower cognitive and
functional level at follow-up than patients with no SSD, even
after adjusting for illness severity, baseline cognitive status,
and severity of baseline functional status (Cole et al., 2003).
In addition, patients with SSD had the same set of risk fac-
tors that predict the likelihood of developing DSM-defined
delirium (e.g., older age, baseline dementia, and greater physi-
ological severity of illness). A recent ICU study similarly con-
cluded that patients with SSD (which defined SSD as patients
who have <4 points on the Intensive Care Delirium Screening
Checklist [ICDSC]) (Bergeron, Dubois, Dumont, Dial,  &
Skrobik, 2001) experienced greater mortality (10.6% vs. 2.4%;
P < 0.001), longer ICU length of stay (5.2 days vs. 2.5 days; P <
0.001), and longer overall hospital stay (40.9 days vs. 31.7 days;
P = 0.002); conversely, patients with no delirium were more
likely to be discharged home and less likely to need convales-
cence or long-term care than those with SSD (see Table 41.8)
(Ouimet et al., 2007).
DI AG N O S I N G DE L I R I U M
Several studies have demonstrated that hospital staff in gen-
eral, and physicians in particular, are not good at identifying
delirium. In fact, studies have demonstrated that delirium is
misdiagnosed, detected late, or missed in over 50% of cases
across the various healthcare settings (Kean & Ryan, 2008).
Despite its high prevalence in critically ill patients, delirium
remains unrecognized in as many as 66%–84% of patients
experiencing this complication (Francis et al., 1990; Inouye,
1994; Rolfson, McElhaney, et al., 1999; Ely, Margolin et al.,
2001; Inouye et al., 2001; Pisani, Redlich et al., 2003; Pisani,
Inouye et al., 2003; Pandharipande et al., 2007; Steis & Fick,
2008; Swigart, Kishi, Thurber, Kathol,  & Meller, 2008;
Steis, Shaughnessy,  & Gordon, 2012). Unfortunately, the
associated mental status changes are often misattributed
to dementia, depression, or just an expected occurrence of
medical illness. Several studies have demonstrated a high
rate (45%–66%) of missed diagnosis (i.e., either missed com-
pletely or misattributed to another disorder) of delirium by
general medicine and surgical services (Farrell  & Ganzini,
1995; Armstrong, Cozza,  & Watanabe, 1997; Kishi et  al.,
2007). A recent study of an adult, acute hospital population
found that only 63.6% of patients with delirium were recog-
nized by nursing staff to be confused or delirious, whereas
43.6% had confusion documented in the medical case notes
(Ryan et al., 2013). The authors concluded that “even if some

41. D E L I R I U M • 857
 Table 41.8 SUBSY NDROMAL DELIR IUM AND CLINICAL OUTCOMES

NO DELIR IUM (ND) SUBSY NDROM AL (SD) CLINICAL(CD) P VALUE*

ICU Mortality 2.4% 10.6% 15.9% P < 0.001

ICU LOS 2.5 d 5.2 d 10.8 d P < 0.001

Hospital LOS 31.7 d 40.9 d 36.4 d ND vs. SD, P = 0.002

Severity of illness
(APACHE II)
*Pair-wise comparison
Source: Ouimet et al., 2007
ND vs. CD, P < 0.001
SD vs. CD, P = 0.137
12.9 16.7 18.6 ND vs. SD, P < 0.001
ND vs. CD, P < 0.001
SD vs. CD, P < 0.016

recognized, but undocumented, delirium cases occurred,
these findings suggest that delirium is not a high diagnostic
and therapeutic priority, despite its treatability and relevance
to outcomes, especially for poorer prognosis in the elderly”
(Ryan et al., 2013).
Factors contributing to the poor detection rate of delirium
can be divided into three large components, which have sig-
nificant interplay (Table 41.9):
1. Patient factors (e.g., older subjects, patients experiencing
3. Systems factors (e.g., lack of consensus over the optimal
comorbid dementia, fluctuating course of presentation,
presence of hypoactive features) (Liptzin & Levkoff,
1992; Inouye et al., 1993; Farrell & Ganzini, 1995;
Inouye, 1999; Inouye et al., 2001; Pisani et al., 2003;
Peterson et al., 2006; Kiely et al., 2007; Lowery et al.,
2007; McAlpine et al., 2008; Meagher et al., 2008;
Givens, Jones, & Inouye, 2009; Leonard et al., 2009;
Collins, Blanchard, Tookman, & Sampson, 2010;
Khurana et al., 2011; Srinonprasert et al., 2011; Morandi,
McCurley, et al., 2012; Steis & Fick, 2012; Ryan
et al., 2013);
2. Clinician/practitioner factors (e.g., lack of knowledge and
training, lack of confidence, lack of suspicion, lack of time
of the clinical staff, expectation that altered mental status
or delirium are a “normal occurrence” in certain medical
settings, such as the ICU) (Wong, Chiu, & Chu, 2005;
Wong, Holroyd-Leduc, et al., 2010; Adams, 1988; Inouye,
1991; Ely, Siegel, & Inouye 2001; Breitbart et al., 2002;
Rabinowitz, 2002; Carnes, Howell, et al., 2003; Cole
et al., 2003; Pisani et al., 2003; Dantzer, Capuron, et al.,
2008; Steis & Fick, 2008; Davis & MacLullich, 2009;
Khan et al., 2009; Collins et al., 2010; Boot, 2012; Steis
et al., 2012); and
assessment of delirium, location of care [worse in surgical
than in medical settings], busy clinical settings [espe-
cially low nurse-to-patient ratio], inadequate application
of “sedation holidays” in sedated-ventilated patients, and
the rapid transfer of patients from one unit to another,
which may decrease the proper documentation and
diagnosis) (Inouye, 1991, 1998; Inouye, Schlesinger, &
Lydon, 1999; Shinn & Maldonado, 2000; Inouye
et al., 2001; Pisani, Redlich, McNicoll, Ely, & Inouye,
2003;(Pisani, Redlich et al. 2003) Siddiqi & House,
2006; Bickel et al., 2008; Davis & MacLullich, 2009;
Collins et al., 2010; NICE, 2010; Wong et al., 2010;
Boot, 2012; Brummel, Morandi, et al., 2012; Neto,
Nassar, et al., 2012; Steis et al., 2012).

Table 41.9 FACTORS CONTR IBUTING TO THE POOR DETECTION R ATE OF DELIR IUM

PATIENT FACTORS CLINICI AN/PR ACTITIONER FACTOR S SYSTEMS FACTORS

• Older subjects • Lack of knowledge and training • Lack of consensus over the optimal assessment
• Patients experiencing • Lack of confidence of delirium
comorbid dementia • Lack of suspicion • Location of care (worse in surgical rather than
• Fluctuating course of • Lack of time of the clinical staff medical settings)
presentation • Expectation that altered mental status or • Busy clinical settings (especially low
• Presence of hypoactive delirium are a “normal occurrence” in certain nurse-to-patient ratio)
features medical settings, such as the ICU • Inadequate application of sedation holidays in
sedated-ventilated patients
• The rapid transfer of patients from one unit
to another, which may decrease the proper
documentation and diagnosis

858 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
 In order to accurately diagnose delirium, clinicians need
to be vigilant and have a high level of suspicion, particularly
in patient populations at higher risk, such as the elderly and
those with the risk factors included in the mnemonic “end
acute brain failure” (see Table 41.2).
To date, the diagnostic gold standard for delirium remains
the Diagnostic and Statistical Manual for Mental Disorders
(DSM). The previously published version was the Fourth
Edition-Text Revised, published in 1994 (DSM-IV-TR; APA,
1994). A  new DSM-5 (5th Edition) has changed the criteria
slightly. The new DSM criteria change the diagnostic focus
away from a “disturbance of consciousness” (DSM-IV) to a
“disturbance in attention”—Criterion A. Criterion B relates
to the acute or subacute onset of symptoms (to differenti-
ate from more chronic cognitive deficits such as dementia;
formerly Criterion C). New Criterion C refers to additional
disturbances in cognition not better explained by another
neurocognitive disorder (akin to former Criterion B in
DSM-IV). New to DSM-5 is Criterion D, which stipulates
that A and B “do not occur in the context of severely reduced
level of arousal or coma.” And, finally, Criterion E is identi-
cal to Criterion D in DSM-IV, which stipulates that there
is physical evidence that can give an organic etiology for the
syndrome. Notice that DSM-5 seems to have got rid of the
presence of psychotic symptomatology (formerly Criterion B
in DSM-IV) as part of the diagnostic criteria.
The International Statistical Classification of Diseases and
Related Health Problems (ICD-10; see Box 41.3) provides its
own diagnostic criteria (WHO, 1992, 2010). Yet, both cri-
teria have at their core the same three cardinal features: dis-
turbance of consciousness (i.e., reduced clarity of awareness
of the environment, with reduced ability to focus, sustain,
or shift attention); global disturbance of cognition (e.g.,
problem-solving impairment or memory impairment) and/
or perceptual disturbance (e.g., illusions or hallucinations);
and an acute or subacute onset with a tendency to fluctuate
(WHO, 2010; APA, 2013). These criteria are commonly used
by medical personnel for the definite diagnosis of delirium.
There are numerous clinically available instruments (see
Box 41.4) developed to help medical providers screen for the
presence of delirium. These instruments were designed to
help nonpsychiatrists (e.g., nurses, internists, and research
assistants) diagnose and follow the progression of delirium
(Trzepacz, Baker,  & Greenhouse, 1988; Marcantonio et  al.,
1994). All of these scales have been derived from, and vali-
dated against, expert psychiatric assessments using DSM or
ICD diagnostic criteria. Some of these tools were designed
and are better used for the purpose of screening for delirium
(e.g., CAM, based on DSM-III-R, original sample size n =
56 [Inouye et al., 1990], CAM-ICU, based on CAM, origi-
nal sample size n = 38 [Ely  et  al.,  2001]). The CAM and
CAM-ICU offer a binary result (i.e., delirium or no delirium),
and there is no severity scale. Unfortunately, these tools have
a high false-positive rate (as high as 10%); therefore, the team
that developed the instrument recommends that all patients
identified as delirious by screening instruments “have fur-
ther evaluation to confirm the diagnosis” (Inouye et  al.,
1990; Liptzin & Levkoff, 1992). In fact, some have expressed
41. D E L I R I U M
Box 41.4 OBJECTIVE MEASUR ES FOR THE DIAGNO-
SIS OF DELIRIUM
• DSM-IV-TR (“Gold Standard”; APA, 1994)
• Delirium Rating Scale (DRS) (Trzepacz, Baker, et al., 1988)
• Confusion Rating Scale (CRS) (Williams, Ward, et al., 1988)
• Confusion Assessment Method (CAM) (Inouye et al., 1990)
• Delirium Symptom Interview (DSI) (Albert et al., 1992)
• Delirium Assessment Scale (DAS) (O’Keeffe, 1994)
• Cognitive Test for Delirium (CTD) (Hart et al., 1996)
• NEECHAM Confusion Scale (Neelon & Champagne, 1996)
• Confusional State Evaluation (CSE) (Robertsson, Karlsson,
et al., 1997)
• Memorial Delirium Assessment Scale (MDAS) (Breitbart
et al., 1997)
• Delirium Severity Scale (DSS) (Bettin, Maletta, et al., 1998)
• Delirium Index (DI) (McCusker, Cole, et al., 1998)
• Delirium Rating Scale–Revised-98 (DRS-R-98) (Trzepacz
et al., 2001)
• Intensive Care Delirium Screening Checklist (ICDSC)
(Bergeron et al., 2001)
• Confusion Assessment Method for the Intensive Care unit
(CAM-ICU) (Ely, Margolin, et al., 2001)
• Delirium Detection Score (DDS) (Otter, Martin, et al., 2005)
• Nursing Delirium Screening Scale (Nu-DESC) (Gaudreau et
al., 2005a)
• Delirium Detection Tool–Provisional (DDT-pro) (Kean,
Trzepacz, et al., 2010)
• Diagnostic and Statistical Manual of Mental Disorders, 5th
Edition (DSM-5) (“New Gold Standard”) (APA, 2013)
concerns regarding the generalizability of the CAM-ICU
findings (Young & Arseven, 2010; Boot, 2012):
On closer scrutiny of the methodology of CAM-ICU
studies, it is important to highlight some consider-
ations when interpreting the data:  (1)  three of the
studies examined were conducted by the tool’s author
who may have had a vested interest, and given that
only 39% of the study population were intubated, it
is difficult to conclude CAM-ICU’s validity in intu-
bated patients (van Eijk, van Marum, et  al., 2009);
(2)  the study population sizes were relatively small,
thus CAM-ICU adaptability into different health-
care organizations remains unclear (Page, Navarange,
et  al., 2009); (3)  examination of the exclusion crite-
ria (e.g., history of psychosis, learning disabilities,
• 859
 depression, dementia) of these studies raises ques-
tions to CAM-ICU transferability to these popula-
tions, especially as these conditions can have similar
characteristics but very different treatment pathways;
(4)  studies have demonstrated although CAM-ICU
has high validity in the non-intubated ICU patients,
it seems to fail to detect milder, subtler symptoms of
hypoactive delirium; (5)  individuals with a history
Box 41.5 PRIMITIVE R EFLEXES
Primitive reflexes are clinical features that indicate brain dysfunc-
tion but that cannot be precisely localized or lateralized. When
present, these signs suggest cortical disease, especially frontal
cortex, resulting in disinhibition of usually extinguished or sup-
pressed primitive reflexes. Their clinical significance is uncertain
and is difficult to correlate with psychiatric illnesses and other
behavior disorders, including delirium.
• Glabellar Reflex: With the examiner’s fingers outside
of patient’s visual field, tap the glabellar region at a rate
of one tap per second. A pathological response is either
absence of blink, no habituation, or a shower of blinks.
Normal response = blinking to the first few taps with rapid
habituation.
• Rooting Reflex: Tested by stroking the corner of the patient’s
lips and drawing away. Pursing of the lips and movement of
the lips or head toward the stroking is a positive response.
• Snout Reflex: Elicited by tapping the patient’s upper lip with
finger or percussion hammer causing the lips to purse and the
mouth to pout.
• Suck Reflex: Tested by placing your knuckles between the
patient’s lips. A positive response would be puckering of
the lips.
• Grasp Reflex: Elicited by stroking the patient’s palm toward
fingers or crosswise while the patient is distracted, causing the
patient’s hand to grasps the examiner’s fingers.
• Palmomental Reflex: Test by scratching the base of the
patient’s thumb (noxious stimulus of thenar eminence).
A positive response occurs when the ipsilateral lower lip and
jaw move slightly downward, and does not extinguish with
repeated stimulation.
• Babinski Sign: Downward (flexor response) movement of the
great toe in response to plantar stimulation.
• Adventitious Motor Overflow: Seen as the examiner tests one
hand for sequential finger movements, and the fingers of
the other hand wiggle or tap. Also if there are choreiform
movements.
• Double Simultaneous Stimulation Discrimination: Tested with
the patient’s eyes closed. The examiner simultaneously brushes
a finger against one of the patient’s cheeks and another finger
against one of the patient’s hands, asking the patient where he
has been touched.
860 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
of mental illness (e.g., psychosis, learning disabilities,
depression, dementia) were excluded, despite reports
indicating delirium being of a medical cause (Boot,
2012, p.187))
In fact, a prospective study was used to assess the diagnostic
validity of the CAM administered at the bedside by nurses
in daily practice during a 5-month period to all patients
(n = 258) consecutively admitted to an acute geriatric ward
of a university hospital (Lemiengre, Nelis, et al., 2006). The
CAM as administered by bedside nurses had a 66.7% sensi-
tivity and a 90.7% specificity. The study suggested that bed-
side nurses had difficulties with the identification of elderly
patients with delirium and that additional education about
delirium is warranted, with special attention to guided train-
ing of bedside nurses in the use of an assessment strategy such
as the CAM for the recognition of delirium symptoms.
A subsequent study compared the diagnostic accuracy of
“delirium experts who diagnosed delirium in 75/181 sub-
jects vs. the CAM-ICU as performed in ‘routine practice’ ”
(i.e., nonresearch setting, conducted by a nonresearch nurse),
where delirium was diagnosed in 35 out of 181 cases. This
yielded a sensitivity of 47% (95% CI = 35%–58%), specificity
of 98% (95% CI  =  93%–100%), positive predictive value of
95% (95% CI = 80%–99%), and negative predictive value of
72% (95% CI  =  64%–79%). This suggested that the “speci-
ficity of the CAM-ICU as performed in routine, daily prac-
tice appears to be high but sensitivity low. The low sensitivity
hampers early detection of delirium by the CAM-ICU” (Van
Eijk et al., 2011).
Moreover, a recent systematic review and meta-analysis
of studies on delirium in critically ill patients (i.e., inten-
sive care units, surgical wards, emergency rooms) published
between 1966 and 2011 (including 16 studies, n = 1,523; and
covering five delirium screening tools) demonstrated that the
CAM-ICU was the most specific bedside tool for the assess-
ment of delirium in critically ill patients (Neto et al., 2012).
Considering only the two most commonly used tools, the
pooled sensitivities and specificities of the CAM-ICU for
detection of delirium were 75.5% and 95.8%, compared to the
ICDSC at 80.1% and 74.6%, respectively. All but one study
was performed in a research setting, and that study suggested
that, with routine use of the CAM-ICU, half of the patients
with delirium were not detected. The authors concluded that
“the low sensitivity of the CAM-ICU in routine, daily prac-
tice may limit its use as a screening test” (Neto et al., 2012).
Other scales have been designed to diagnose and assess
the severity of a delirium episode. These include the Delirium
Rating Scale (DRS), based on DSM-III, original sample size
n = 20 (Trzepacz et al., 1988); the DRS-98, based on DSM-IV,
original sample size n = 68 (Trzepacz, Mittal, et  al., 2001);
the Memorial Delirium Assessment Scale (MDAS), based
on DSM-IV, original sample size n = 68 (Breitbart et  al.,
1997); and the Intensive Care Delirium Screening Checklist
(ICDSC), based on DSM-IV, original sample size n = 93
(Bergeron et  al., 2001), among others. (Please see Box 41.4
for a comprehensive list of delirium diagnostic tools.) A sig-
nificant advantage of diagnostic tools that measure delirium
severity is that they provide clinicians a means to measure
the severity of the episode and determine whether the con-
dition seems to be worsening or improving (i.e., consecutive
measures). A severity score is also useful in research studies, as
particular interventions may not be able to fully prevent delir-
ium but may lessen the severity of the episode. Finally, sever-
ity scales may provide the ability to diagnose subsyndromal
delirium (i.e., patients presenting with mental status changes
that do not rise to the level of full DSM or ICD diagnostic
criteria). Clinical experience and accumulating research data
seem to indicate that the presence of subsyndromal symp-
toms may have similarly negative effects in the patients who
are experiencing them. A study of ICU patients demonstrated
that patients experiencing subsyndromal delirium experi-
enced significantly greater ICU mortality (10.6% vs. 2.4%),
longer ICU stays (5.2 days vs. 2.5 days), and longer total hospi-
tal stays (40.9 days vs. 31.7 days) compared to patients experi-
encing similar medical illness but without delirium (Ouimet
et al., 2007). As expected, these subsyndromal patients expe-
rienced less ICU mortality (10.6% vs. 15.9%), shorter ICU
stays (5.2  days vs. 10.8  days), but longer total hospital stays
(40.9  days vs. 36.4  days) compared to patients experiencing
similar medical illness but with full clinical delirium (see
Table 41.8).
In general, no delirium screening or assessment tool is
expected to be perfect. Some have criticized the fact that even
the two psychometrically valid scales most commonly used in
the ICU (i.e., CAM-ICU, ICDSC) have detection ranges of
delirium from 10% to >80%, in similar populations. It may
very well be that vigilance and close clinical monitoring are
more useful than any commonly used tool. In fact, some
authors have reported that the CAM-ICU scoring may be
affected by sedation, thus limiting its diagnostic accuracy
(Kress, 2010). A study of “real-life” conditions (as opposed to a
research setting) suggested that “the CAM-ICU in daily prac-
tice showed not quite as good test characteristics as presented
in the original validation studies” (van Eijk et  al., 2011). In
fact, these authors found that “after stratification according to
type of delirium, sensitivity of the CAM-ICU was lowest in
the hypoactive subgroup (31%; 95% CI, 17%–48%); highest
in the hyperactive delirious patients (100%; 95% CI, 56%–
100%); and intermediate in the mixed-type patients (53%;
95% CI, 35%–74%); exhibiting particularly poor test char-
acteristics in neurocritical care patients (sensitivity 17%; 95%
CI, 1%–64%) (van Eijk et al., 2011). A recent study suggested
that clinical assessments by critical care physicians may iden-
tify delirium more rapidly and more accurately than screening
tool assessments (Bigatello et al., 2013).
It is unclear what the new DSM-5 diagnostic criteria will
mean to the diagnostic accuracy of existing tools. It is likely
that some of them may require revisions, yet  all will need
to be validated taking the new criteria into consideration.
Nonetheless, it may be important for the clinician to be aware
that the 2013 American College of Critical Care Medicine
(ACCM)/Society of Critical Care Medicine (SCCM) clini-
cal practice guidelines for pain, agitation, and delirium
(PAD), based on available evidence, strongly recommend that
critically ill patients be routinely monitored for delirium in
41. D E L I R I U M
the ICU using a validated tool. After their review of avail-
able delirium assessment tools, the 2013 PAD guideline
group concluded that the Confusion Assessment Method
for the ICU (CAM-ICU) and the Intensive Care Delirium
Screening Checklist (ICDSC) are the ICU delirium screen-
ing tools with the strongest validity and reliability (Davidson,
Harvey, Bemis-Dougherty, Smith, & Hopkins, 2013).
The most critical part of the assessment, given the char-
acteristic waxing and waning of this syndrome, is to add to
the clinical examination all available clinical information,
including interview of the family members and caregiv-
ers, nursing and medical staff, and a thorough review of the
chart for behaviors exhibited during the preceding 24 hours.
Another potential clue of the presence of delirium may come
from a thorough neurological examination. In our clinical
experience, patients with delirium tend to exhibit a reemer-
gence of primitive signs (see Box 41.5). This appears to be
more consistent in cases of hypoactive delirium. The relation-
ship between poor cognitive status and primitive reflexes has
been described in patients suffering from HIV-related cogni-
tive disorders (Tremont-Lukats, Teixeira, et al., 1999) and in
cases of dementia (Paulson, 1977). One study described the
presence of primitive reflexes in post-cardiotomy patients
suffering from postoperative neuropsychiatric complica-
tions (Liu & Hsieh, 1993). A second study found that 55% of
patients exhibiting at least two primitive reflexes and 80% of
those exhibiting at least three primitive reflexes in the postop-
erative period developed delirium (Nicolson, Chabon, et al.,
2011). Further studies are needed to determine whether an
assessment for the presence of primitive reflexes may add to
the diagnostic accuracy for delirium or at least assist in the
characterization of delirium type, and whether such assess-
ment has any prognostic value.
Some have advocated the use of the electroencephalogram
(EEG) as a way to identify and diagnose delirium. Engel and
Romano were the first to describe the relationship between
delirium and the diffuse slowing and progressive disorganiza-
tion of rhythm seen in the EEG (Engel, Webb, & Ferris, 1945)
(Engel, Webb et al. 1945). The most common EEG findings in
delirium include slowing of peak and average frequencies, and
decreased alpha activity but increased theta and delta waves.
Studies suggest that EEG changes correlate with the degree
of cognitive deficit, but there does not appear to be a rela-
tionship between EEG patterns and delirium motoric type
(Romano & Engel, 1944; Engel & Romano, 1959; Koponen,
Partanen, et al., 1989; Koponen, Hurri, et al., 1989; Trzepacz,
Sclabassi, & Van Thiel, 1989; Trzepacz, Leavitt, et al., 1992;
Engel  & Romano, 2004; Plaschke, Hill, et  al., 2007). The
clinical usefulness of EEG in the diagnosis of delirium may
be constrained by its limited specificity (given that there are
many conditions and medications that may affect the EEG)
and the practicality of conducting the test (particularly in the
case of agitated and combative patients). Still, the EEG can
be useful in differentiating delirium from other psychiatric
and neurological conditions such as catatonic states, seizure
activity (e.g., nonconvulsive status), medication side effects
(e.g., posterior reversible encephalopathy syndrome due to
the use of calcineurin inhibitors), or the manifestations of
• 861
the behavioral and psychological symptoms associated with
dementia (BPSD). BPSD affects between 50% and 90% of
patients suffering from dementia and is usually characterized
by three main syndromes:  agitation/aggression, psychosis,
and an affective component. At times, apathy (as in the case of
hypoactive delirium) or disinhibition (as in the case of hyper-
active delirium) can be seen, particularly when there is frontal
lobe involvement (Finkel, Costa e Silva, et al., 1996).
Finally, a 24-hour accelerometer-based activity monitor
has been used to categorize the motoric behavior of patients
with delirium. The continuous wavelet transform (CWT)
provided by the instrument can then be used to character-
ize a delirium as hyperactive, hypoactive, or mixed (Godfrey,
Conway, et  al., 2009, 2010; Meagher, 2009). The technique
has yet to find a clinical use, however.
M A N AG E M E N T OF   DE L I R I U M
A DE QUAT E M E DIC A L M A N AG E M E N T
The management of delirium has five main components (see
Box 41.6):
1. Recognition of patients at risk (see Tables 41.1
and 41.2);
2. Implementation of prevention techniques (with pharma-
cological and nonpharmacological approaches), especially
in populations identified to be at high risk;
3. Enhanced surveillance and screening;

Box 41.6 ALGORITHM FOR THE MANAGEMENT OF DELIRIUM

I. Recognition of patients at risk

A. A particular patient’s odds of developing delirium are associated with the interaction between the following conditions:
i. Knowledge of a patient’s characteristics.
ii. Predisposing and precipitating medical risk factors (Table 41.4).
iii. Modifiable and non-modifiable risk factors for that particular patient or patient population (Table 41.5).
iv. Specific medical conditions and surgical procedures the patient is exposed to.
B. Obtaining the patient’s baseline level of cognitive functioning using information from accessory sources (e.g., Informant
Questionnaire on Cognitive Decline in the Elderly [IQCODE]).
II. Implementation of prevention techniques
A. A key focus should be placed on prevention strategies, particularly in “at risk” populations.
i. Avoid all pharmacological agents with high deliriogenic potential or anticholinergic load, if possible.
ii. Promote a non-pharmacological sleep protocol.
iii. Early mobilization.
B. For patients in the ICU, especially those on ventilation or IV sedation, consider:
i. Sedating to a prescribed or target sedation level (e.g., RASS).
ii. Consider using the sedative agent with lowest deliriogenic potential:
• Dexmedetomidine use is associated with the lowest incidence of delirium.
• Propofol use is a good second choice; followed by midazolam.
iii. Reassess pain levels daily and titrate opioid agents to the lowest effective level required to maintain adequate analgesia.
• Hydromorphone is preferred as baseline agent of choice for pain management.
• Limit the use of fentanyl to rapid initiation of analgesia and as rescue agent.
• Avoid the use of opioid agents for sedation or management of agitation or delirium.
iv. Provide daily sedation holidays:
• Interrupt sedative infusions daily until the patient is awake
• Restart sedation, if needed, at the lowest effective dose
• Reassess target sedation level (e.g., RASS).
III. Enhanced surveillance, screening and early detection
A. Most important aspects of surveillance:
i. Knowledge about the condition and presenting symptoms.
ii. A high level of suspicion.
B. Be vigilant for the development of delirium in high-risk groups:
i. Use a standardized surveillance tools (e.g., DRS-R-98; MDAS; CAM)
ii. Use psychiatric consultants (i.e., DSM-5/ICD-10 criteria)
iii. Be particularly aware of the presence of hypoactive delirium and its different manifestations.
C. Use psychiatric consultants to help with assessment and design of the treatment plan, if available.
D. Train medical personnel at all levels.
IV. Treatment of all forms of delirium
A. Identify and treat underlying medical causes.
(continued)

862 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
 B. Treat or correct underlying medical problems and potential reversible factors.
C. Conduct an inventory of all pharmacological agents that have been administered to the patient.
i. Discontinue any medication or agent known to cause delirium or to have high anticholinergic potential, if possible, or insti-
tute a suitable alternative.
D. Implement early mobilization techniques, to include ALL of the following components:
i. Daily awakening protocols (sedation holiday), as described above.
ii. Remove IV lines, bladder catheters, physical restraints, and any other immobilizing apparatuses as early as possible.
iii. Initiate aggressive PT and OT as soon as it is medically safe to do.
a. In bedridden patients, therapy may be limited to daily passive range of motion.
b. Once medically stable, get the patient up and moving as early as possible.
iv. Provide patients with any required sensory aids (i.e., eyeglasses, hearing aids).
v. Promote as normal a circadian light rhythm as possible.
a. Better if this can be achieved by environmental manipulations, such as light control (i.e., lights on and curtains open
during the day; lights off at night) and noise control (i.e., provide ear plugs, turn off TVs, minimize night staff chatter).
b. Provide as much natural light as possible during the daytime.
vi. Provide adequate intellectual and environmental stimulation as early as possible.
E. Avoid using GABA-ergic agents to control agitation, if possible.
• Exception: cases of CNS-depressant withdrawal (i.e., alcohol, benzodiazepines, barbiturates); or when more appropriate
agents have failed and sedations are needed to prevent patient’s harm.
F. Adequately assess and treat pain.
i. Avoid the use of opioid agents for behavioral control of agitation.
ii. Rotate opioid agents from morphine to hydromorphone or fentanyl.
G. For the treatment of delirium (all types), consider using:
i. Acetylcholinesterase inhibitor (e.g., rivastigmine) for patients with a history of recurrent delirium or delirium superimposed
on known cognitive deficits. Physostigmine, for known causes of anticholinergic delirium.
ii. Melatonin (e.g., 3 mg q HS) or melatonin agonists (e.g., ramelteon 8 mg q HS) to promote a more natural sleep. If that is
ineffective, consider trazodone (e.g., 25–100 mg q HS) or mirtazapine (e.g., 3.75–7.5 mg q HS).
iii. Serotonin antagonist (e.g., ondansetron).
H. In case of hyperactive delirium, consider the use of the following agents:
i. Dopamine antagonist agents (to address DA excess) (e.g., haloperidol, risperidone, quetiapine, aripiprazole).
a. Moderate-dose haloperidol (e.g., <20 mg/24 hr in divided doses), is still considered the treatment of choice, if the
patient’s cardiac condition allows it.
b. Before using haloperidol:
• Obtain 12-lead ECG; measure QTc.
• Check electrolytes; correct K+ and Mg+, if needed.
• Carefully review the patient’s medication list and identify any other agents with the ability to prolong QTc.
• If possible, avoid other medications known to increase QTc and/or inhibitors of CPY3A4.
c. When the use of haloperidol is contraindicated or not desirable, atypical antipsychotics should be considered:
• There is better evidence for risperidone and quetiapine.
• ii.Limited data for: olanzapine, aripiprazole, perospirone.
• Avoid: clozapine, ziprasidone.
d. Discontinue dopamine antagonist agents’ use if QTc increases to >25% of baseline or >500 msec.
ii. Alpha-2 agonist agents (to address the NE excess) (e.g., dexmedetomidine, clonidine, guanfacine)
a. Consider changing primary sedative agents from GABA-ergic agents (e.g., propofol or midazolam) to an alpha-2 agent
(e.g., dexmedetomidine), starting at 0.4 mcg/kg/hr, then titrate dose every 20 minutes to targeted RASS goal.
iii. Anticonvulsant and other agents with glutamate antagonism or Ca+ Ch modulation (e.g., VPA, gabapentin, amantadine,
memantine).
I. Consider the use of NMDA-receptor blocking agents, to minimize glutamate-induced neuronal injury (e.g., amantadine,
memantine), particularly in cases of TBI and CVA.
J. G. In case of hypoactive delirium:
i. Evidence suggests that DA antagonists may still have a place, given the excess DA theory.
a. If haloperidol is used, recommended doses are in the very low range (i.e., 0.25–1.0 mg/24 hr). This is usually given as a
single nighttime dose, just before sundown.
b. If an atypical is preferred, consider an agent with low sedation (i.e., risperidone, aripiprazole).
ii. In cases of extreme psychomotor retardation or catatonic features, in the absence of agitation or psychosis, consider the use
of psychostimulant agents (e.g., methylphenidate, dextroamphetamine, modafinil) or conventional dopamine agonists (e.g.,
bromocriptine, amantadine, memantine).

(Modified from Maldonado, 2008a, 2009, 2011)

4. Treatment or correction of underlying medical problems
and potentially reversible factors (e.g., correction of
electrolyte imbalances, infection, end organ failure,
sleep deprivation, pain, metabolic and endocrinological
disturbances, substance or medication intoxication or
withdrawal); and
5. Treatment of all forms of delirium (with pharmacological
and nonpharmacological approaches).
We advocate an integrative approach that incorporates
input and collaboration among all pertinent medical teams,
including psychosomatic medicine, critical care, internal
medicine, surgery, neurology, nursing, respiratory therapy,
physical therapy, and occupational therapy teams, which we
like to call the “CNS pharmacotherapy algorithm” (see Figure
41.22) (Maldonado, 2008a, 2011, 2014). The idea is to opti-
mize patient comfort (i.e., adequate sedation), minimize pain
(i.e., adequate analgesia), prevent or treat delirium, restore an
adequate sleep–wake pattern, begin physical mobilization as
early as possible, improve cognitive recovery, and return to
baseline functional level as soon as medically possible. Others
have found psychiatric consultations to be superior to a mul-
tidisciplinary geriatric team approach in treating individuals
with depression and delirium (Cole, Fenton, et al., 1991).
The recognition of patients at risk begins with knowl-
edge of your patient’s characteristics; an assessment of the
predisposing and precipitating medical risk factors to which
your patient is, or may be, exposed (Table 41.2); and being
acquainted with the modifiable and nonmodifiable risk fac-
tors for that particular patient or patient population (Table
41.3). Finally, certain medical conditions and surgical proce-
dures are more likely to be associated with the development
of delirium than others. It is likely that a particular patient’s
odds of developing delirium are associated with the interac-
tions among these conditions. These have been discussed in
detail elsewhere (Maldonado, 2008a, 2014).
Prevention techniques have been found to be rather effec-
tive, especially when targeted to patients at high risk for devel-
oping delirium. There is a whole host of pharmacological and
nonpharmacological techniques available. It is important
that providers carefully consider the potential side effects ver-
sus the benefits associated with effective delirium prevention.

Is Pt @ targeted RASS?
Default RASS: -1 to +1
NO
YES Reassess Daily Goals

YES
Are there known Continue Daily
intracranial processes? Awakening Protocol

Neurology Consult
&
Management RASS
Algorithm Score?

RASS= +2 to +3 RASS= ≥+4

RASS= -2 to -5

YES
Re-assess Metabolic problems? Address & Correct This constitutes a
Neuro Medico/Psychiatric
Status Emergency
NO

YES Initiate Pain

Hold/DC ALL Pain? Management
CNS depressant Rxs Algorithm
Dose opioids to pain level; not NO
behavior-reassess 1hr
Initiate Substance
Likely YES Assess pertinent Physiological causes:
Abuse & With drawal
Substance Abuse
Management Algorithm • Metabolic
Or Withdrawal?
• Pain
Improved MSE? • Anxiety
NO • Delirium
YES
Assess & Initiate • CNS Stimulant Abuse
YES Anxiety Management • CNS Depressant Withdrawal
Likely Anxiety
Algorithm
Hypoactive Disorder?
Delirium
NO
If possible harm to patient/staff, initiate
Psych Consult & Initiation Psych Consult & Initiation
Behavioral Rescue Protocol/
Hypoactive Delirium Hyperactive Delirium Call urgent Psychiatry Consult
Algorithm Algorithm

Figure 41.22 CNS Pharmacotherapy Model. (Adapted from Maldonado, 2008a, 2011, 2014)

864 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
The early recognition and prompt treatment of delirium is
of utmost importance, not only because of issues of safety
for patient and staff, but also because of the serious negative
consequences of no treatment or delayed treatment. In fact,
a recent study demonstrated that patients whose delirium
treatment was delayed were more frequently mechanically
ventilated (50.0% vs. 22.3%; p = 0.012), had more nosoco-
mial infections (including pneumonia) (p < 0.05), and had
a higher mortality rate (p < 0.001) (see Figure  41.23) than
patients whose treatment was promptly started (Heymann
et  al., 2010). These data suggest that a delay in initiating
delirium therapy may be associated with increased morbidity
and mortality.
The most important aspects of adequate surveillance
and early detection include knowledge about the condition
and presenting symptoms (of all motor forms) and main-
taining a high level of suspicion, especially in populations
at risk. There are several surveillance tools and techniques
(discussed earlier in this chapter) that can be used to effi-
ciently screen subjects, especially those at greatest risk.
Adequate training of medical personnel at all levels is para-
mount. Surveillance should be effectively implemented by
all practitioners.
Once diagnosed, adequate management of delirium
includes the following steps:
1. Management of the behavioral and psychiatric manifesta-
tions and symptoms to prevent the patient from self-harm
or harming of others (e.g., use of tranquilizing agents to
manage agitation);
1.0
Delayed therapy
Immediate therapy
0.8
Cumulative survival
0.6
0.4
0.2
0.0
0 20 40 60
Time since first occurence of delirium (days)
Survival Estimates for Delirious ICU Patients with
Figure 41.23
Immediate or Delayed Delirium Treatment. Kaplan–Meier survival
estimates for 184 ICU patients with delirium treated within 24 hours of
diagnosis (immediate therapy group) and 20 patients in whom delirium
treatment started > 24 hours after diagnosis (delayed therapy group);
the delayed therapy group showed a lower probability of surviving until
ICU discharge than the immediate therapy group (vertical markers
show censorship points). There was a significantly higher risk of death in
the delayed therapy group versus the immediate therapy group (log rank
test, P < 0.001; Breslow test, P = 0.013). (Heymann et al., 2010)
41. D E L I R I U M
2. Treatment or correction of underlying medical problems
and potentially reversible factors (e.g., correction of electro-
lyte imbalances, infection, end organ failure, sleep depriva-
tion, pain, metabolic and endocrinological disturbances,
substance or medication intoxication or withdrawal); and
3. Correction of the neurochemical derangement (triggered
by the underlying cause) which leads to the behavioral
manifestations of delirium (see Table 41.7).
NON PH A R M AC OL O G IC A L
M A N AG E M E N T S T R AT E G I E S
Nineteen studies have been published on nonpharmacological
prevention strategy in the non-ICU environment (see Table
41.10). The original study suggesting that nonpharmacological
approaches may be of use was a nonrandomized clinical trial,
where hospitalized elderly patients were assessed for manifes-
tations of delirium in response to the correction of environ-
mental factors commonly associated with increased risk for
delirium (Inouye et al., 1999). The multicomponent protocol
consisted of simple techniques applied by the hospital staff,
including reorientation, appropriate cognitive stimulation
three times a day, the implementation of a nonpharmacologi-
cal sleep protocol to help normalize patient’s sleep–wake cycle,
early mobilization after surgery or extubation, timely removal
of catheters and restraints, correction of sensory deficiencies
(i.e., by eyeglasses and hearing aids), and early correction of
dehydration and electrolyte abnormalities. As a result of these
environmental manipulations, they observed an astonishing
40% reduction in the risk of delirium (see Figure 41.24). This
landmark work was preceded by two negative studies and fol-
lowed by seven positive and five negative ones. Among the
positive studies, there were three positive randomized clinical
trials, and these are condensed in Table 41.11. They were all
conducted among elderly orthopedic patients, and all found
that the implementation of nonpharmacological multicom-
ponent interventions led to a reduction in the occurrence of
delirium and fewer medical complications in the intervention
groups (Marcantonio, Flacker, 2001; Vidan, Serra, Moreno,
Riquelme, & Ortiz, 2005; Lundstrom et al., 2007). Given the
findings reported by others, a multicomponent approach has
been developed that targets identified, treatable contributing
factors and addresses them early on: the Hospital Elder Life
Program (Inouye  & Charpentier, 1996; Inouye, Bogardus,
et al., 1999). This program involves the implementation, by an
interdisciplinary team, of targeted interventions for known
80 risk factors (i.e., cognitive impairment, sleep deprivation,
immobility, dehydration, vision or hearing impairment) for
cognitive decline in the elderly.
More recently, a meta-analysis of published studies has
found that, although the multicomponent interventions
appeared to be effective in preventing delirium among post-
operative hip replacement patients, the intervention made
no difference in a number of important secondary measures,
including discharge location or post-discharge dependency,
length of hospital stay (p = 0.12), or mortality rate (p = 0.77)
between intervention and control groups (Holroyd-Leduc,
• 865
Table 41.10 NONPHAR MACOLOGICAL PR EVENTION STR ATEGY STUDIES IN THE NON-ICU EN VIRONMENT

DELIR IUM INCIDENCE % P VALUE

DELIR IUM
STUDY(N = 19) POPULATION INTERVENTION DEFINITION CONTROL INTERVENTION

Schindler et al., 1989 CABG NP—peri-op Psych intervention vs. DSM-III 0 (0/17) 12.5 (2/16) ns
RCT, n = 33 usual care

Wanich et al., 1992 Gen IM elderly NP—nursing intervention for elderly DSM-III 22 (22/100) 19 (26/135) p = 0.61
NRCT, n = 235 patients hospitalized patients vs. usual care

Inouye et al., 1999 Gen IM elderly NP—multi-component intervention CAM 15 (64/426) 9.9 (42/426) p = 0.02
NRCT, n = 852 patients vs. usual care

Millisen et al., 2001 Traumatic hip Fx Sx NP—multi-component vs. usual care CAM 23.3 (14/60) 20 (12/60) p = 0.82
NRCT, n = 120 repair

Marcantonio et al., 2001 Elderly pts. after Hip NP—multi-component intervention CAM 50 (32/64) 32 (20/62) p = 0.04
RCT, n = 126 Fx Sx vs. usual care

Tabet et al., 2005 Gen IM elderly NP—Staff education vs. usual care Single assessment 19.5 (25/128) 9.8 (12/122) p = 0.034
NRCT, n = 250 patients psychiatrist

Wong et al., 2005 Traumatic hip NP—multi-component vs. usual care CAM 35.7 (10/28) 12.7 (9/71) p = 0.012
Pre- and post-eval Fx Sx repair

Vidan et al., 2005 Elderly pts. after Hip NP—multi-component intervention CAM 45.2 (70/155) 61.7 (100/164) p = 0.003
RCT, n = 319 Fx Sx vs. usual care For ≥1 major
complications

Lundstrom et al., Elderly pts. after NP—multi-component intervention Organic Brain 75.3 (73/97) 54.9 (56/102) p = 0.003
RCT, n = 199 Hip Fx Sx vs. usual care Syndrome
Scale (OBSs)

(Caplan & Harper, 2007) Geriatric ward NP—usual care vs. CAM 38.1 (8/21) 6.3 (1/16) p = 0.032
Pre–post eval, n = 37 volunteer-mediated intervention
(Inouye style)

Taguchi et al., 2007 Esophageal CA Normalization of their natural NEECHAM scale 40 (2/5) 16 (1/6) P = 0.014
RCT, n = 11 subjects circadian rhythm by light therapy

Benedict, Hazelett, Acute Care for Elders NP—Delirium prevention protocol Modified [3.24] [3.76] p = 0.368
et al., 2009 (ACE) units vs. usual care NEECHAM scale
NRCT, n = 65

Schweickert et al., 2009 MICU Early exercise and mobilization (PT & CAM-ICU Return to Return to pp = 0.0202
RCT, n = 104 OT) at daily sedation interruption vs. independence independence by
sedation interruption by D/H: 35% D/H: 59%
Delirium Delirium duration:
duration: 2 days
4 days
Holroyd-Leduc Traumatic hip NP—multi-component delirium CAM Pre-implemen- Post-implemen- p = 0.84
et al., 2010 Fx Sx repair strategies tation tation incidence
NRCT, n = 134 incidence 31
33 (20/64)
(23/70)

Holroyd-Leduc Comprehensive geriatric assessment CAM (2) The multicomponent

et al., 2010 and multi-component Organic Brain interventions for
Meta-analysis, 3 studies interventions targeted at precipitants Syndrome the management of
(n = 489) of delirium Scale (1) delirium had:
• No effect on mortality
(summary RR 1.08,
95% CI 0.81–1.44;
p for heterogeneity =
0.77).
• No effect on length
of stay (summary
weighted mean
difference 3.25 days,
95% CI -2.85 to
9.34 days; p for
heterogeneity =0.12)
• There was no impact
on post-discharge
dependency,
function, or the need
for institutional care.

Björkelund, Hommel, Elderly hip Fx Sx NP—multi-component delirium OBS Scale 34 (45/132) 22 (29/131) p = 0.096
et al., 2010 repair strategies
NRCT, n = 263

Colombo et al., 2012; All patients admitted NP—patients underwent both CAM-ICU 35.5 (60/170) 22 (31/144) p = 0.020
n = 314; to mixed (med/surg) a reorientation strategy and
ICU over a year environmental, acoustic, and visual
stimulation. Study compared pts.
in observation arm (control) vs.
intervention arm

Gagnon et al., 2012, Palliative care NP—multi-component administered Confusion rating 43.9 (370/842) 49.1 (330/674) P = 0.045
randomized delirium patients, in 2 cancer to patient and family education vs. scale (CRS)
prevention trial; n = 1516 centers usual care
Martinez et al., 2012 Older adults in gen Randomized to receive a CAM 13.3 (19/143) 5.6 (8kal/14 4) P = 0.027
n = 287 medicine ward multi-component management
protocol, delivered by family
members (144 patients) or standard
management (143 patients).
 0.25
Cumulative Incidence
0.20
of Delirium
0.15
0.10
0.05
0.00
1 3 5 7
Day
A MultiComponent Intervention to Prevent Delirium in
Figure 41.24
Hospitalized Older Patients. (Inouye et al., 1999)
Khandwala, & Sink, 2010). A separate study on the same pop-
ulation and by the same researcher, but this one conducted
in a nonresearch environment, found that the intervention
had no effect on the overall delirium rate (p = 0.84) and
no significant difference on secondary measures (i.e., mean
length of hospital stay), falls, or discharge to long-term care
facilities (Holroyd-Leduc et  al., 2010). Finally, a large study
(n = 1,516) followed patients from admission to death at
seven palliative care centers and compared usual care versus
a nonpharmacological delirium-preventive intervention. The
study found that the multicomponent preventive intervention
had no effect on the primary outcome, overall delirium inci-
Usual care dence (p = 0.66); and, similarly, it had no effect on secondary
outcomes of delirium severity, total delirious days, or dura-
tion of the first delirium episode (Gagnon, Allard, Gagnon,
Merette, & Tardif, 2012).
Finally, a study examined the administration of nonpro-
Intervention fessional, family-provided multicomponent intervention
among high-risk elderly patients admitted to a medicine ward
Median length
(Martinez, Tobar, Beddings, Vallejo,  & Fuentes, 2012). The
of stay intervention consisted of educating family members about
the clinical features and prognostic implications of delirium,
9 11 plus the usual content of the traditional multicomponent
intervention, but fully administered by family members. They
found a significant reduction in the occurrence of delirium in
the intervention group (p = 0.027). It is unclear why the mul-
ticomponent protocol yielded no results when implemented
by the nursing staff in a “real world” scenario, but a possible
explanation is that the average medicosurgical unit may not
have the same nurse-to-patient ratio as research units, making
the implementation of the protocol challenging. Therefore, an
approach that combines the best of all available protocols and
that takes into consideration available resources and limita-
tions of each clinical setting may work best.
Of significance, a study of mechanically ventilated, critically
ill patients used a combination of early physical and occupa-
tional therapy during periods of daily interruption of sedation
to demonstrate a significantly greater return to independent
functional status (p = 0.02), shorter duration of delirium
(p = 0.02), and more ventilator-free days (p = 0.05) (Schweickert

Table 41.11 SELECTED NONPHAR MACOLOGICAL RCT-PR EVENTION INTERVENTION STUDIES

STUDY AUTHORS;
TYPE; SIZE STUDY DETAILS FINDINGS

Marcantonio Studied the effectiveness of proactive
et al., 2001 geriatric consultation compared to usual
RCT, n = 126 care in reducing delirium in patients 65
and older admitted emergently for surgical
repair of hip fracture.
While there were not statistical differences between intervention
and control groups regarding baseline measures and characteristics,
the study found a reduction in the occurrence of delirium in the
intervention group (32%) compared to usual care (50%) (P = 0.04).
There was an even greater reduction in cases of “severe delirium,”
occurring in 12% of intervention patients and 29% of usual-care
patients. Despite this reduction in delirium, length of stay did not
significantly differ between intervention and usual care.

Vidan et al., 2005 Studied patients were randomly assigned
RCT, n = 319 to daily multidisciplinary geriatric
intervention or usual care during
hospitalization in the acute phase of hip
fracture.
Lundström Studied randomly assigned elderly patients
et al., 2007 after femoral neck fracture repair to
RCT, n = 199 postoperative care in a specialized geriatric
ward (i.e., staff education focusing on the
assessment, prevention, and treatment of
delirium and associated complications) or a
conventional orthopedic ward.
The study showed that intervention subjects experienced shorter
hospital stay (median length 16 vs. 18 days) (P = 0.06); lower
in-hospital mortality (0.6% vs. 5.8%, P = 0.03) and lower rate of major
medical complications (45.2% vs. 61.7%, P = 0.003). Overall, after
adjustment for confounding variables, the geriatric intervention
was associated with a 45% lower probability of death or major
complications. More patients in the geriatric intervention group
achieved a partial recovery at 3 months (57% vs. 44%, P = 0.03),
but there were no difference between groups at 6 and 12 months
assessments.
Patients in the intervention group experience less postoperative
delirium (54.9% vs. 75.3%, p = 0.003); those who did experience
delirium did so for a shorter duration compared with controls (5.0 ±
7.1 days vs. 10.2 ± 13.3 days, p = 0.009). Patients in the intervention
group suffered from fewer complications (e.g., decubitus ulcers,
urinary tract infections, nutritional complications, sleeping problems,
and falls) and shorter hospital stays (28.0 ± 17.9 days vs. 38.0 ±
40.6 days, p = 0.028).

868 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
et al., 2009). This is the first randomized, nonpharmacological
intervention study shown to reduce ICU delirium. Others have
found similar results (i.e., lower total amount benzodiazepine
use, higher level of functional mobility, and reduction in delir-
ium rates) by combining sedation reduction and early mobi-
lization with physical rehabilitation (Needham & Korupolu,
2010). It is likely that the minimization of offending agents
(e.g., sedatives), along with early mobility, may have something
to do with the decreased rate of delirium among this cohort.
Previously, I have suggested an approach that incorporates
nonpharmacological and pharmacological approaches to the
prevention and treatment of delirium, which is summarized in
Box 41.6 but expanded elsewhere (Maldonado, 2008a, 2009,
2011). Given the significant adverse consequences of delirium,
the primary goal should be prevention, particularly in popu-
lations at high risk. Key among them is minimizing the det-
rimental effects of prolonged and uninterrupted sedation.
Therefore, we recommend that sedation be targeted to a pre-
determined goal using sedation scales such as the Richmond
Agitation Sedation Scale (RASS; see Sessler, et al., 2002). As
previously noted, daily awakening and sedation holidays are
highly recommended. These will allow patients to awaken;
then, if still needed, sedation should be restarted at the lowest
needed level—again targeting a predetermined sedation level.
Certain sedative agents are preferred to others, given pharma-
cological and pharmacodynamic qualities; and some appear to
have less deliriogenic potential than others (Maldonado, van
der Starre, Wysong, & Block, 2004; Pandharipande, Pun et
al., 2007; Maldonado, 2008a, 2009, 2011; Riker et al., 2009).
Therefore, dexmedetomidine may be the preferred choice
when clinically indicated and tolerated, followed by propofol,
and last by midazolam. Similarly, some analgesic agents may
be better at preventing or controlling delirium. For example,
despite its faster onset of action, fentanyl’s clearance is much
slower than hydromorphone; thus it tends to accumulate, and
this may be problematic (see Figure 41.25). Therefore, we pre-
fer the use of hydromorphone as the baseline agent of choice
for pain management and sedation supplementation and limit
the use of fentanyl for rapid initiation of analgesia and as a res-
cue agent.
More recently, a randomized study suggested that not
using sedation, compared to the standard treatment with
sedation and a daily wake-up trial, may further reduce the
300
yl
an
Fe
nt
Minutes to 50% decrement
240
in plasma concentration
180 Meperid
ine
120
tha
Me
60
Sufentanil
0
0 120 240 360
Infusion Duration
Minutes to 50% decrement in plasma concentration of com-
Figure 41.25
monly used opioid agents.
time patients require mechanical ventilation, shorten the
length of the patient’s stay in the intensive care unit, and
decrease the overall hospital length of stay (Strom  & Toft,
2011), thus providing further evidence that the use of seda-
tive drugs should be reduced. Patients should be mobilized
as early as possible, and individual patient sedative needs
should be evaluated on a daily basis to minimize the untow-
ard effects of these agents and optimize the care of each indi-
vidual patient.
Given the high rate of underdiagnosed and missed cases,
the routine use of assessment scales or diagnostic interviews
by properly trained personnel will facilitate timely treatment
and prevention of complications of delirium. Early involve-
ment of a medical-psychiatric team is valuable for both pre-
vention and early intervention. An active search for possible
etiologies of delirium must first rule out the common causes
of the syndrome (see Table 41.2). This must include a review
of all medications and identification, and possible discon-
tinuation of agents with high deliriogenic potential (see
Table 41.4). Appropriate diagnostic tests and assays should
be ordered and reviewed in a timely fashion and all abnormal
findings addressed accordingly.
Conduct an inventory of all pharmacological agents that
have been administered to the patient. Any medication or
agent known to cause delirium (see Table 41.4) or to have high
anticholinergic potential (see Table 41.5) should be discontin-
ued if possible; if not, a suitable less anticholinergic alternative
should be substituted.
Immobilizing lines and devices (e.g., chest tubes, IV lines,
bladder catheters) should be removed as early as possible.
Similarly, physical restraints should be avoided and elimi-
nated as soon as it is safe to do. More importantly, early mobi-
lization (e.g., daily interruption of sedation, minimization
of restraint use, early physical and occupational therapy) has
proven to be a key factor in minimizing delirium and improv-
ing the odds of returning to independent functioning upon
discharge home (Schweickert et al., 2009).
Early correction of sensory deficits should be under-
taken. That is, eyeglasses and hearing aids should be replaced
or fitted (if the patient was not using them before the hos-
pitalization) as soon as possible. This will allow patients to
familiarize themselves with the environment and reorient
themselves early on. It will also minimize the occurrence
of misperceptions or misinterpretation of environmental
cues and stimuli. Environmental isolation should be mini-
mized if possible. Family members and loved ones should be
encouraged to visit and provide a familiar and friendly envi-
ronment, as well as provide appropriate orientation and stim-
ulation to patients, especially those with baseline cognitive
deficits. Dehydration and electrolyte abnormalities should
e
don be corrected as quickly and safely as possible. Malnutrition
Alfentanil
Morphine should be corrected, unless the patient is end-stage and their
Remifentanil
Hydromorphone
delirium is a pre-terminal phenomenon. In that situation,
480 600
the decision to use a feeding tube, for example, would be a
personal decision that different patients and families would
make differently.
For intubated, sedated ICU patients, the approach
must include judicious management of sedative and pain
41. D E L I R I U M • 869
management agents in order to allow for early extubation, PH A R M AC OL O G IC A L M A N AG E M E N T
minimize time in bed, and limit the amount of potentially S T R AT E G I E S
deliriogenic agents a patient is exposed to. Daily cessation of
It cannot be overstated that the “definitive treatment” of
sedative protocols combined with daily spontaneous breath-
delirium is the accurate identification and treatment of its
ing trials and early mobilization (through physical and occu-
underlying cause(s). Nevertheless, pharmacological interven-
pational therapy) have resulted in a significant decrease in the
tion with various psychoactive agents is often needed to help
duration of mechanical ventilation, the length of stay in inten-
manage agitated patients and for the correction of the neu-
sive care, and the number of days patients experienced acute
rotransmitter derangements associated with delirium symp-
brain dysfunction, compared with control groups (Kress,
toms. Of note, no pharmacological agent has received Food
Pohlman, O'Connor, & Hall, 2000; Kress et al., 2003; Girard
and Drug Administration (FDA) approval for the treatment
et  al., 2008; Schweickert et  al., 2009; Needham, Korupolu,
of delirium.
et al., 2010).
In an attempt to “first, do no harm,” we should first
Finally, environmental manipulations to normalize the
identify all agents that may contribute to, exacerbate, or
sleep–wake cycle may be considered, from reducing noise lev-
perpetuate delirium. Chief among these are anticholinergic
els and decreasing nighttime tests and procedures to increas-
substances and GABA-ergic agents (Box 41.2). As described
ing the amount of natural light during daytime hours. Early
above, all such agents (e.g., benzodiazepines, propofol) may
correction of sleep disturbance, preferably by nonpharmaco-
cause or aggravate delirium and its behavioral manifestations
logical means, should be attempted.
(Marcantonio et  al., 1994; Ely et  al., 2001; Pandharipande,
Some have advocated the use of bright light therapy
Shintani, et  al., 2006). The use of benzodiazepines in the
(similar to that used for the treatment of seasonal affec-
management of delirium should be limited to (a)  patients
tive disorder) as an alternative to medication agents but
experiencing delirium related to the withdrawal from a
despite the logic behind this approach, very little evidence
CNS-depressant agent (e.g., alcohol, barbiturates, benzodiaz-
for it exists. A  small pilot study demonstrated that bright
epines), or (b) when other more appropriate agents have failed
light therapy (5000 lux applied for 2 hours in the morning)
and the level of agitation and need for behavioral control out-
led to a significant difference (P = 0.014) in the incidence
weighs the potentially detrimental effects of benzodiazepines.
of delirium (Taguchi, Yano, & Kido, 2007). Yet there were
Following the same principle, we should avoid the use of opi-
serious methodological problems with this study, including
oid agents for behavioral control of agitated patients and use
a very small sample, lack of blinding, and late initiation of
these agents only for pain management, as opioids have been
the intervention (after extubation rather than immediately
implicated in the development of delirium in many patient
after surgery). A  second, larger study (n = 228) conducted
populations (Morrison et al., 2003; Centeno, Sanz, & Bruera,
in a geriatric monitoring unit compared bright light therapy
2004; Morita et  al., 2005; Fong et  al., 2006; Vaurio et  al.,
(2000–3000 lux; 6–10 pm daily) against placebo. The study
2006; Gaudreau et al., 2007, Vella-Brincat & Macleod, 2007;
found that among those randomized to bright light therapy,
Wang et al., 2007). A summary of a comprehensive delirium
there were significant improvements in modified Barthel
management algorithm can be found in Figure 41.26.
Index (MBI) scores, especially for the hyperactive and
Pharmacological management strategies have mostly been
mixed delirium subtypes (P < 0.05); significant improve-
limited to the use of antipsychotic agents, acetyl cholinester-
ments on the DRS sleep–wake disturbance subscore, for
ase inhibitors, and sedative-hypnotic agents. In this section,
all delirium subtypes; and improvements in the mean total
I  will address possible pharmacological interventions based
sleep time (from 6.4 hours to 7.7) (P < 0.05) and length of
on the neurotransmitter being targeted. In general, deliri-
first sleep bout (SB; 6.0 compared with 5.3 hours) (P < 0.05),
ous patients have an excess of dopamine, norepinephrine,
with decreased mean number of SBs and awakenings when
glutamate, and hyperactivity of calcium channels (Ca+ Ch);
compared to those in the placebo group (Chong, Tan, Tay,
reduced levels of acetylcholine and melatonin; and variable
Wong, & Ancoli-Israel, 2013).
levels of serotonin, histamine, and gamma-aminobutyric
The National Institute for Health and Clinical Excellence
acid, depending on the type and cause of delirium. We will
(NICE) provided a set of guidelines containing 13 specific
tailor the discussion regarding pharmacological prophylaxis
recommendations for the prevention of delirium in elderly
and treatment based on these assumptions. Studies will be
at-risk patients (O’Mahony, Murthy et  al., 2011). They are
discussed in order of publication. In the only published set of
mostly based on the correction of modifiable factors that may
guidelines for sedation management of critically ill adults, the
precipitate delirium, and the implementation of the multi-
Society of Critical Care Medicine recommended haloperidol
component intervention package. The full version of these
as “the preferred agent for the treatment of delirium in criti-
recommendations can be found at their website: http://guid-
cally ill patients” (Jacobi, Fraser, et  al., 2002). Accordingly,
ance.nice.org.uk/CG103/Guidance/pdf/English. Similarly,
haloperidol is used by about 78%–80% of intensivists as the
the Society of Critical Care Medicine has developed pain,
standard pharmacological agent to manage delirium in the
agitation, and delirium guidelines, promoted mobility to
ICU, while novel agents, the second-generation antipsychot-
improve care of critically ill patients, and has developed
ics, are gaining popularity for delirium management, are now
tools to facilitate and rapidly implement the translation of
being used by 35%–40% of ICU physicians (Patel, Gambrell,
guideline care recommendations into practice (Davidson
et al., 2009; MacSweeney, Barber, et al., 2010). A more recent
et al., 2013).

870 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
 COMPREHENSIVE DELIRIUM ALGORITHM

Delirium Suspected

Non-Rx Delirium Treatment - For ALL Deliria Types:

! Correct malnutrition, dehydration & electrolyte abnormalities
Consider intracranial, (“–“) ! Remove non-essential immobilizing devices & lines
CAM/
metabolic, noxious & ! Review Rx lits and remove potential offending agents
CAM-ICU
psychiatric causes
! Correct sensory deficits
(“+“) ! Implement non-Rx sleep protocol to promote circadian rhythm
(“–“) ! Provide adequate intellectual stimulation– avoid TV
DRS-R-98 ! Address pain- but avoid the use of opioids to control behavior
! Avoid all GABAergic agents & anticholinergic substances
(“+“) ! Continue daily awakening protocol on intubated/sedated patients
Neurology Consult Psychiatry Consult Initiate Pain
& Management & Management Management
Algorithm Algorithm Algorithm Delirium
Management
Protocol Hypoactive Delirium

Initiate Substance
Intoxication or Assess for Substance
Withdrawal Intoxication or
Management (“+“) Withdrawal
Delirium Hypoactive Delirium Algorithm:
Algorithm ! Avoid GABA-ergic & anticholinergic agents
Motoric Type
(“–“) AND
! Add Melatonin 6 mg PO q 2000; or ramelteon 8 mg q 2000
Initiate Anxiety ! Consider ultra-small dose haloperidol (if QTc <500 msec)
Assess for Anxiety (“–“)
Management ! If SGA, use low dose risperidone or aripiprazole (if QTc <500 msec)
(“+“) (GAD-7) ! Consider use of rivastigmine
Algorithm
! In cases of extreme psychomotor retardation consider the use
(“–“) of psychostimulant agents
Monitor
Initiate Depression Hypoactive Delirium ! Obtain daily EKG- D/C antipsychotics if QTc > 500 msec
Assess for
Management ! Use of other Rx that may prolong QTc
(“+“) Depression (PHQ)
Algorithm
(“–“)

Initiate Psychosis Hyperactive Delirium Algorithm:

Management Assess for Psychosis ! Avoid GABA-ergic & anticholinergic agents
Algorithm (“+“)
AND
! Add Melatonin 6 mg PO q 2000; or ramelteon 8 mg q 2000
! Consider the use of alpha-2 agonist agents for sedation/agitation
! Consider moderate dose haloperidol (if QTc <500 msec)
! If SGA, use low dose risperidone or quetiapine (if QTc <500 msec)
! Consider use of rivastigmine, If Hx of cognitive deficit or dementia, or in protracted cases
! In cases of extreme psychomotor retardation consider the use of psychostimulant agents
Monitor
! Obtain daily EKG-D/C antipsychotics if QTc > 500 msec
! Use of other Rx that may prolong QTc

Figure 41.26 Comprehensive Delirium Algorithm. Source: (Maldonado, 2008a, 2011)

“post-publication perspective” of practice guidelines added no
new information about pharmacological agents for the man-
agement of delirium in the ICU, but emphasized the benefits
of early detection and patient mobilization. It is important
to point out that soon after the publication of these perspec-
tives a number of significant studies were published, all con-
clusively demonstrating the delirium-prophylactic benefits of
antipsychotic agents in at-risk populations, which will be dis-
cussed in the next section.
Of note, no pharmacological agent has received FDA
approval for the prevention of delirium.
Dopamine Excess
Dopamine Antagonists for Delirium Prevention
There are very few control studies on the efficacy of antipsy-
chotic agents for the prevention of delirium, yet most studies
suggest a positive effect (see Table 41.12). The first random-
ized controlled trial compared haloperidol to placebo in
at-risk, elderly orthopedic patients (Kaneko, Jianhui, et  al.,
1999). The incidence of delirium was significantly lower in
the treatment group (10.5% vs. 32.5%; p <.05). Similarly, the
intensity and duration of postoperative delirium were more
severe and lasted longer in the control group.
The first double-blind, randomized, placebo control stud-
ies (DBRPCT) compared haloperidol to placebo, also in
at-risk elderly orthopedic patients (Kalisvaart, de Jonghe,
et al., 2005). Subjects received oral haloperidol 0.5 mg/d up to
72 hours preoperatively and then until the third postoperative
day. The study found that prophylactic haloperidol use did not
alter the incidence of postoperative delirium when compared
with placebo (15.1 vs.16.5%; p = ns). Nevertheless, it did have
a positive effect on the severity (the mean highest DRS-R-98
score ±SD was 14.4 ± 3.4 vs. 18.4 ± 4.3; P <.001) and dura-
tion (5.4 vs. 11.8 days; P <.001) of delirium and shortened the
length of hospital stay (mean number of days in the hospital
was 17.1±11.1 vs. 22.6 ± 16.7; P <.001).
Many additional studies have followed, all with positive
findings. A  study of patients undergoing cardiac surgery
with cardiopulmonary bypass (CPB), randomly assigned to
receive either 1 mg of risperidone or placebo sublingually
as soon as they regained consciousness postoperatively,

41. D E L I R I U M • 871
Table 41.12 ANTIPSYCHOTIC AGENTS FOR PR EVENTION OF DELIR IUM

DELIR IUM INCIDENCE%

STUDY(N = 9) POPULATION INTERVENTION DELIR IUM DEFINITION CONTROL/INTERVENTION P-VALUE

Kaneko et al., 1999; RPCT GI surgery Prophylaxis haloperidol vs. PBO IV DSM-III-R 32.5 10.5 p < 0.05
postoperatively for 5 days

Kalisvaart et al., 2005 Elderly hip-replacement Sx PBO vs. haloperidol 1.5 mg/d started DSM-IV 16.5 15.1 p = 0.91
DBRPCT, n = 430 preop, continued for up to 3 days postop. CAM (36/216) (32/212)
DRS-R98

Prakanrattana & Cardiac Sx under CPB PBO vs. sublingual risperidone CAM 31.7 11.1 p = 0.009
Prapaitrakool, 2007 immediately p-Sx (20/63) (7/63)
DBRPCT, n = 126

Girard et al., 2010* Med/Surg ICU in PBO vs. haloperidol vs. ziprasidone: days CAM-ICU 12.5 14.0 15.0 p = 0.66
DBRPCT, n = 101 mechanical ventilation alive without delirium or coma; [1.2–17.2] [6.0–18.0] [9.1–18.0]
conducted in 6 tertiary medical centers. days days days

Larsen et al., 2010 Elderly elective total PBO vs. 5 mg of orally disintegrating DSM-III-R 40.2 14.3 p < 0.001
DBRPCT, n = 495 joint-replacement olanzapine 1-dose pre- and 1-dose (82/204) (28/196)
post-surgery

Wang et al., 2012 Elderly, non-cardiac Sx PBO vs. HAL (0.5 mg bolus, followed by CAM 23.2 15.3 p = 0.031
DBRPCT, n = 457 cont infusion 0.1 mg/hr x 12hrs)

van den Boogaard et al., ICU patients at “high risk” PBO vs. HAL (1 mg/8 hr.) within 24 hrs. CAM-ICU 75 65 P = 0.01
2013, 2013)DBRPCT, n = of admission to ICU
476

Hirota & Kishi, 2013; Various clinical settings Meta-analysis of 6 studies (3 HAL, Various tools Sensitivity analysis showed that SGA P < 0.00001
Meta-analysis (RCTs), 6 1 olanzapine, 2 risperidone) using were superior to PBO (NNT = 4; p <
studies, n = 1689 antipsychotic agent for delirium 0.0001), whereas HAL failed to show
prophylaxis superiority to PBO
Teslyar et al., 2013; Postoperative elderly Medication administered included Various tools The pooled relative risk of the five P < 0.01
Meta-analysis (RCTs); 5 patients haloperidol (3), risperidone (1), and studies resulted in a 50% reduction
studies, n = 1491 olanzapine (1). in the relative risk of delirium
among those receiving antipsychotic
medication compared with placebo.
* Underpowered.
Abbreviations: RPCT=randomized placebo control trial; GI=gastrointestinal; PBO=placebo; DBRPCT=double blind randomized, placebo controlled trial; DSM=Diagnostic and Statistical Manual of Mental Disorders;
CAM= Confusion Assessment Method; DRS=Delirium Rating Scale; Sx=surgery; CPB=cardio-pulmonary by-pass machine; ICU=intensive care unit. APA=antipsychotic agent; OLA=olanzapine; HAL=haloperidol;
RIS=risperidone
demonstrated a significantly lower incidence of postop-
erative delirium in the risperidone group (11.1% vs. 31.7%;
P = 0.009) (Prakanrattana & Prapaitrakool, 2007). A study
examined the effects of olanzapine pretreatment (5 mg
Zydis® formulation administered just preoperatively, and
again 5 mg administered immediately after surgery upon
awakening) to placebo in elderly subjects undergoing ortho-
pedic joint-replacement surgery. The incidence of delirium
in the intervention group was significantly lower in the
study group compared to placebo (15% vs. 41%; P < 0.0001)
(Larsen, 2007).
A large (n = 457; ages ≥65) postoperative delirium pre-
vention study compared the efficacy and safety of low-dose
IV-haloperidol (i.e., 0.5 mg intravenous bolus injection, fol-
lowed by continuous infusion at a rate of 0.1 mg/hr for 12 hrs
vs. placebo) vs. placebo in critically ill patients after noncar-
diac surgery. The study found the incidence of delirium dur-
ing the first 7 days after surgery was 15.3% in the haloperidol
group vs. 23.2% in the control group (p = .031). Other mea-
sures also appeared significantly improved:  the mean time
to onset of delirium and the mean number of delirium-free
days were significantly longer (6.2  days [95% CI 5.9–6.4]
vs. 5.7 days [95% CI 5.4–6.0]; p = .021; and 6.8 ± 0.5 days
vs. 6.7 ± 0.8 days; p = .027, respectively); the median length
of ICU stay was significantly shorter (21.3 hrs [95% CI,
20.3–22.2] vs. 23.0 hrs [95% CI, 20.9–25.1]; p = .024) in the
haloperidol group than in the control group. There was no
significant difference with regard to all-cause 28-day mortal-
ity between the two groups (0.9% [2/229] vs. 2.6% [6/228];
p = .175). No drug-related side effects were documented
(Wang et al., 2012).
Finally, a study assessed the efficacy of haloperidol pro-
phylaxis against delirium in “at risk” ICU patients (n =
177) (van den Boogaard, Schoonhoven, van Achterberg, van
der Hoeven,  & Pickkers, 2013). The results of the interven-
tion were compared to a historical control group and a con-
temporary group that did not receive haloperidol prophylaxis,
mainly due to noncompliance to the protocol, mostly during
the implementation phase. The protocol consisted of low-dose
haloperidol infusion (i.e., IV-haloperidol 1 mg/8 hr (or a
lower dose, for patients ≥80 years), started as soon as patients
were identified to be at increased risk, within the 24-hour
period after ICU admission. All patients experiencing delir-
ium received therapeutic doses of haloperidol according to
the unit’s customary protocol. Patient characteristics were
comparable between the prevention and the control groups;
with the exception of prophylactic haloperidol, as per study
protocol. The intervention resulted in a lower delirium inci-
dence (65% vs. 75%, p =.01), more delirium-free days (median,
20  days [interquartile range 8 to  27] vs. median 13  days
[3  to  27], p =.003), fewer ICU readmissions (11% vs. 18%,
p =.03), and less frequent unplanned removal of tubes/lines
(12% vs., 19%, p =.02) in the intervention group compared to
the control group. Beneficial effects of haloperidol appeared
most pronounced in the patients with the highest risk for
delirium. Haloperidol was stopped in 12 patients because of
QTc-time prolongation (n = 9), renal failure (n = 1) or sus-
pected neurological side-effects (n = 2). No other side effects
41. D E L I R I U M
were reported, including torsades de point (TdP) (van den
Boogaard et al., 2013).
Two meta-analyses of studies using dopamine antagonist
agents for delirium prophylaxis found that pooled relative risk
of published studies suggested a 50% reduction in the relative
risk of delirium among those receiving antipsychotic medica-
tion compared with placebo (p <0.01). The studies suggest
that perioperative use of prophylactic dopamine antagonist
agents (both typical and second-generation antipsychotics),
compared to placebo, may effectively reduce the overall risk
of postoperative delirium, thereby potentially reducing mor-
tality, disease burden, length of hospital stay, and associated
healthcare costs (Hirota & Kishi, 2013; Teslyar, Stock, et al.,
2013). A  third meta-analysis, including all postoperative
prevention strategies, found that both typical (3 RCTs; n =
965; RR = 0.71; 95% CI = 0.54–0.93) and atypical antipsy-
chotics (3 RCTs; n = 627; RR = 0.36; 95% CI = 0.26–0.50)
decreased delirium occurrence, compared to placebos (Zhang
et al., 2013).
Dopamine Antagonists for Delirium Treatment
The literature has long recognized that intravenous neuro-
leptic agents are the recommended emergency treatment for
agitated and mixed type delirium (Adams, Fernandez,  &
Andersson, 1986; Fernandez, Holmes, Adams, & Kavanaugh,
1988; Sanders, Murray,  & Cassem, 1991; Ziehm, 1991;
Riker, Fraser, & Cox, 1994; Inouye et al., 1999). Intravenous
administration of haloperidol has always been thought supe-
rior to oral administration because the IV route has more
reliable absorption, even in cases of systemic organ failure.
Intravenous haloperidol use has the added advantage of
requiring no patient cooperation, thus facilitating its use even
in uncooperative and agitated patients. Studies suggest that
the IV use of high-potency neuroleptic agents is associated
with minimal effects on blood pressure, respiration, and heart
rate (Ayd, 1978; Tesar, Murray,  & Cassem, 1985; Adams,
1988; Sanders et  al., 1991; Riker et  al., 1994; Stern, 1994;
Segatore & Adams, 2001). And, some have suggested that hal-
operidol has a lower incidence of extrapyramidal symptoms
when administered IV versus orally (7.2% vs. 22.6%; p <0.01)
(see Figure  41.27) (Sanders, Minnema,  & Murray, 1989;
Maldonado & Kang, 2003).
Intravenous haloperidol has been identified as the agent of
choice for the management of delirious critically ill patients by a
100%
80%
60% PO
40%
IV
23%
7%
20%
0%
EPS
Haloperidol Route of Administration and Incidence of EPS.
Figure 41.27
Rate of extrapyramidal symptoms and their association to mode of
administration for haloperidol. (Maldonado, 2003)
• 873
number of national organizations, including Britain’s National
Institute for Health and Clinical Excellence (NICE, 2010) the
American Psychiatric Association (APA, 1999, 2004)  and
the Society of Critical Care Medicine (Shapiro et  al., 1995;
Jacobi et  al., 2002). Since then, a “best evidence topic in car-
diac surgery” was written according to a structured proto-
col, addressing the issue of haloperidol safety for critically ill
patients (Khasati, Thompson, & Dunning, 2004). Their search
included 294 papers and concluded that haloperidol should be
considered the first-line drug for agitated patients following
cardiac surgery. Of note, in September of 2007, the FDA issued
a “black-box” warning for the off-label clinical practice of using
IV-haloperidol (FDA, 2007). It is important to remember that
haloperidol has never been approved by the FDA for IV use.
Other antipsychotic agents have been studied for the
treatment of delirium. Most of these include atypical or
second-generation antipsychotic (SGA) agents. Because of
the stigma and potential side effects associated with typical
antipsychotics, SGAs have being been used at increasing rates
over the last few years for the management of psychiatric and
behavioral symptoms (e.g., agitation, psychosis, delirium) in
medically ill patients. Large studies, particularly head-to-head
comparisons between SGAs and more conventional agents
(i.e., haloperidol), are lacking; therefore, most clinical prac-
tice is derived from case reports and open-label (OL) studies.
Following are summaries of the only seven RCTs available; all
others are summarized in Table 41.13.
The first study was a DBRCT that compared haloperidol
(average. dose 1.4 mg/d) vs. chlorpromazine (average dose
36.0 mg/d) versus lorazepam (avg. dose 4.6 mg/d) among
HIV/AIDS patients (Breitbart et al., 1996). Results showed
that treatment with either haloperidol or chlorpromazine in
relatively low doses resulted in significant improvement in the
symptoms of delirium, while no improvement was found in
the lorazepam group. Treatment with either neuroleptic was
associated with an extremely low prevalence of extrapyrami-
dal side effects; all patients receiving lorazepam, however,
developed treatment-limiting adverse effects.

Table 41.13 DOPAMINE ANTAGONIST—ANTIPSYCHOTIC AGENTS AS TR EATMENT OF DELIR IUM

DELIR IUM
DEFINITION/
MEASUR E OF
STUDY(N = 32) POPULATION INTERVENTION R ESPONSE R ESULTS

Breitbart et al., AIDS, medical Haloperidol vs. DSM-IIIR/DRS Tx either HAL or CPM resulted in
1996; DBRCT, patients chlorpromazine vs. significant improvement in the symptoms
n = 30 lorazepam of delirium, while no improvement was
found in the LOR group. Tx neuroleptic was
associated with an extremely low prevalence
of EPS, while all patients receiving LOR
developed treatment-limiting adverse effects.

Sipahimalani & Med/Surg patients Haloperidol vs. DRS Improvement was similar in both groups
Masand, 1998; OL, olanzapine (mean DRS +SD H = 11.1 ± 7.1; O = 10.3 ±
n = 22 4.8; P = 0.760), with extrapyramidal
symptoms found only in haloperidol
patients. No side effects in olanzapine
group.

Schwartz et al., Med/Surg patients Quetiapine vs. DRS Effectiveness of 350% in reducing DRS
2000; Single-blind; haloperidol— scores. When compared with haloperidol,
n = 11 retrospective chart there was no difference in onset of symptom
review resolution, duration of treatment, and
overall clinical improvement.

Kim et al., 2001; Med/Surg patients Olanzapine PO, DSM-IV/DRS 50% decrease in Delirium Rating Scale
OL, n = 20 variable dose scores (from pre of 20.0 ± 3.6, to post
of 9.3 ± 4.6; P < 0.01). No side effects,
including EPS.

Breitbart et al., Hospitalized Olanzapine PO, DSM-IV/MDAS Olanzapine was effective in treating 76%
2002; OL, n = 79 cancer patients variable dose of delirium patients as evidenced by the
MDAS; caused excessive sedation in 30%
of patients.

Horikawa et al., Med/Surg patients Risperidone PO DSM-IV/DRS At a low dose of 1.7 mg/d, on average,
2003; OL, n = 10 risperidone was effective in 80% of
patients, and the effect appeared within
a few days. Most commonly cited adverse
effects included sleepiness (30%) and mild
drug-induced Parkinsonism (10%).
(continued)

874 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
Table 41.13 (CONTINUED)

DELIR IUM
DEFINITION/
MEASUR E OF
STUDY(N = 32) POPULATION INTERVENTION R ESPONSE R ESULTS

Sasaki et al., 2003; Med/Surg patients Quetiapine PO, DSM-IV/DRS 100% of patients on quetiapine achieved
OL, n = 12 flexible doses resolution of delirium (mean on day
4.8 ± 3.5 days); no EPS reported.

Kim et al., 2003; Elderly medical Quetiapine PO, DSM-IV/DRS 100% of patients on quetiapine achieved
OL, n = 12 inpatients flexible doses resolution of delirium by day 10 (mean
on day 5.9 ± 2.2 days); no EPS reported.
Delirium Rating Scale scores along
with scores of the Mini-Mental State
Examination and Clock Drawing Test
continued to improve throughout the
3-month study period.

Liu et al., 2004; Med/Surg patients Risperidone DSM-IV/various Patients treated with haloperidol were
retrospective record with hyperactive (average dose younger than patients treated with risperidone
review, n = 77 delirium 1.17±0.76 mg/d) vs. (P < 0.05). The mean hyperactive syndrome
haloperidol (avg. scale score was higher in the haloperidol
dose 4.25±2.62 than that of the risperidone group. The
mg/d) authors found no significant difference in the
efficacy or frequency of response rate between
haloperidol and risperidone (100% vs. 95%;
p = ns). Patients on risperidone experienced
less EPS (7% vs. 69%).

Mittal et al., 2004; Patients admitted to Risperidone, 0.5 mg DSM-IV/DRS Rapid resolution of delirium while
OL, n = 10 Med-Surg unit PO BID, flexible receiving low dose
PRNs risperidone (mean dose 0.75 mg/d); no EPS
reported.

Parellada, Baeza, Prospective, Risperidone PO DSM-IV/ Risperidone was administered at the time
et al., 2004; OL multicenter, DRS of diagnosis, and treatment was maintained
observational PANSS-P according to clinical response. Found
7-day study MMSE a significant decrease in DRS scores in
90.6% of treated patients and significantly
improved all symptoms measured by the
scales from baseline to day 7 (p < 0.0); only
3% side effects.

Pae, Lee, et al., 2004 Med/Surg patients Quetiapine PO DRS-R9/ DRS-R-98 and clinical global scores were
OL, n = 22 CGI8 significantly reduced by 57.3% and 55.1%,
respectively. Quetiapine was effective
and safe.

Han et al., 2004; Med/Surg patients Haloperidol vs. CAM/ Both groups showed significant improvement
DBRCT, n = 28 risperidone, 7 days DRS in baseline DRS and MDAS scores with
medication trial MDAS either haloperidol (75%) or risperidone (42%)
(P < 0.05). There was no significant difference
in improvement of DRS (p = 0.35) or MDAS
(p = 0.51) scores, comparing haloperidol with
risperidone patients.

Hu et al., 2004; Med/Surg elderly Haloperidol vs. DRS/ All groups showed a decrease in DRS scores
RPCT, n = 175 patients olanzapine vs. CGI by 7th day, compared with baseline (p <
placebo, 7 days 0.01). Decrease in DRS scores of treated
medication trial patients at day 7 (OLA 72.2%; HAL 70.4%)
differed significantly from DRS scores of
PBO patients (29.7%; (p < 0.01), but not
from each other (p > 0.05).
(continued)
Table 41.13 (CONTINUED)
DELIR IUM
DEFINITION/
MEASUR E OF
STUDY(N = 32) POPULATION INTERVENTION R ESPONSE R ESULTS

Skrobik et al., 2004; Critically ill Med/
OL-prospective RCT, Surg patients
n = 73
Haloperidol (avg. DIS ICU Delirium Index Screening Checklist
6.5 mg/d) vs. Scores were reduced in both groups,
olanzapine (avg. 4.5 compared with baseline (p < 0.05), but there
mg/d) was no significant difference in DIS scores
between active Tx groups (p = 0.9). EPS
were found in 13% of haloperidol patients,
but 0% in olanzapine group.

Toda et al., 2005; n Elderly inpatient Risperidone, OL, DSM-IV/ Resolution reported in 7 subjects (mean
= 10 general medicine 0.5 mg oral sol.; DRS dose 0.92 ± 0.47 mg/d); 1 non-responder;
flexible titration, side effects requiring Tx discontinuation
PRN reported in 2.

Lee et al., 2005; Med/Surg patients Amisulpride vs. DRS-R98/ After treatment, DRS-R-98 scores were
RCT, n = 40 quetiapine CGI significantly decreased from the baseline
in both treatment groups (P < 0.001)
without group difference. Both atypical
antipsychotics were generally well tolerated.

Straker et al., 2006 Medically ill patients Aripiprazole PO DSM-IV/ 50% of patients had improved significantly
OL, n = 14 was used in a flexible DRS-R98 by day 5, as indicated by a 50% reduction
dosing range, from CGI in DRS-R-98 scores. 86% of patients had a
5 mg/day to 15 mg/ 50% reduction in their DRS-R- 98 scores
day, titrated as by end of treatment. Mean CGI Severity
clinically indicated scores at the beginning of treatment were
5.2, with a mean CGI Improvement score
after treatment of 2.1, indicating much
improvement.

Takeuchi et al., 2007; Med/Surg patients
OL, n = 38
Perospirone, OL DSM-IV/ Perospirone was effective in 86.8% of
DRS-R98 patients, within several days (5.1 ±
4.9 days). The initial dose was 6.5 ± 3.7 mg/
day and maximum dose of perospirone was
10.0 ± 5.3 mg/day. There were no serious
adverse effects.

Maneeton, Medically ill patients Quetiapine, flexible CAM/DRS, CGI 88% subjects responded. Means (SDs) dose
Maneeton, & dosing and duration (SD) of quetiapine treatment
Srisurapanont, 2007; were 45.7 (28.7) mg/day and 6.5 (2.0) days,
OL, n = 17 respectively. The DRS and CGI-S scores of
days 2–7 were significantly lower than those
of day 0 (p < 0. 001) for all comparisons).
Only two subjects were shown to have mild
tremor.

Reade et al., 2009; Med/Surg ICU Agitated delirium ICDSC DEX significantly shortened median
OL-RCT were randomized to Time time to extubation from 42.5 to 19.9 hours
receive an infusion (P = 0.016); it significantly decreased
of either haloperidol ICU length of stay, from 6.5 to 1.5 days
0.5 to 2 mg/hour or (P = 0.004); and of patients requiring
dexmedetomidine ongoing sedation, it reduced the time
0.2 to 0.7 μg/kg/hr propofol was required in half (79.5% vs.
41.2%; P = 0.05).

Devlin et al., 2010; MICU PBO vs. quetiapine ICDSC Tx w QUE was associated with: a shorter
DBRPCT, n = 36 (50 mg BID) time to first resolution of delirium
(multienter-3) (p =.001), a reduced duration of delirium
(p =.006), less agitation (p =.02), greater
chance to be discharged home vs.
long-term-care facility (p =.06), and lower
requirement of as-needed haloperidol
(p =.05).
(continued)
Table 41.13 (CONTINUED)

DELIR IUM
DEFINITION/
MEASUR E OF
STUDY(N = 32) POPULATION INTERVENTION R ESPONSE R ESULTS

Girard et al., 2010; Mechanically PBO vs. HAL vs. CAM-ICU Patients in the haloperidol group spent
DBRPCT, n = 101 ventilated Med/Surg Ziprasidone a similar number days alive without
ICU patients delirium or coma (14.0 [6.0 –18.0] days) as
did those on ziprasidone (15.0 [9.1–18.0]
days) and PBO groups (12.5 [1.2–17.2] days;
p = 0.66).

Kim et al., 2010; Elderly, Med/Surg Risperidone vs. DSM-IV/DRS-r-98 Significant within-group improvements
SB-RCT, n = 32 patients olanzapine in the DRS-R-98 scores over time were
observed at every time point in both
treatment groups. The response rates
did not differ significantly between the
two groups (risperidone group: 64.7%,
olanzapine group: 73.3%); and there were
no difference in the safety profiles and side
effects between groups.
Tahir et al., 2010; Med/Surg patients Quetiapine vs. DSM-IV/DRS-r-98, Quetiapine has the potential to more
DBRCT, n = 42 placebo CGI quickly reduce the severity of non-cognitive
aspects of delirium. The study was
underpowered for treatment comparisons.

Grover et al., 2011; Med/Surg patients Haloperidol DSM-IV/DRS-r-98 Patient in all three groups experience a
Prospective, single (0.25 to 10 mg) significant reduction in DRS-R98 severity
blind, n = 64 vs. olanzapine scores and a significant improvement in
(1.25 to 20 mg) vs. MMSE scores over the period of 6 days;
risperidone (0.25 to with no difference between the treatment
4), flexible dosing groups. Rate of side effects was also similar.

Boettger et al., 2011; Med/Surg patients at Aripiprazole, DSM-IV/MDAS The mean dosage of aripiprazole required
OL, n = 21 Cancer Center flexible dosing was 18.3 mg (range of 5–30) daily at
T3. Patients treated for delirium with
aripiprazole experienced significant
improvement and resolution of delirium,
with MDAS scores declining from a mean
of 18.0 at baseline (T1) to mean of 10.8 at
T2 and a mean of 8.3 at T3. There was a
100% resolution of hypoactive delirium,
compared to patients with hyperactive
delirium (58.3%).

Hakim et al., 2012; Patients aged Randomized using a ICDSC Seven (13.7%) patients in the risperidone
PCRCT 65 yrs. or older computer-generated group experienced delirium vs. 17
who experienced list to receive (34%) in the placebo group (p = 0.031).
subsyndromal placebo (n = 50) or Competing-risks regression analysis showed
delirium after 0.5 mg risperidone that failure to treat subsyndromal delirium
on-pump cardiac (n = 51) every 12 hrs with risperidone was an independent
surgery by mouth risk factor for delirium (p = 0.002). Two
(3.9%) patients in the risperidone group
experienced extrapyramidal manifestations
versus one (2%) in the placebo group
(p = 1.0).

Kishi et al., 2012; Adult delirious Risperidone given DRS-R98 Entry DRS-R-98 score = 19.8 ± 6.8;
OL, n = 29 cancer patients orally once per day 7-d follow-up score = 14.3 ± 7.8, results
(mean dosage, 1.4 ± demonstrate DRS-R98 scores were
1.3 mg/day improved in 79.3% of patients during the
study period (p < 0.001); 38% achieved
remission (i.e., DRS-R-98 ≤10).
(continued)
Table 41.13 (CONTINUED)

DELIR IUM
DEFINITION/
MEASUR E OF
STUDY(N = 32) POPULATION INTERVENTION R ESPONSE R ESULTS

Tagarakis et al., Consecutive patients Ondansetron IV (8 Self-developed rating Noted a statistically significant
2012; n = 80 who developed mg) vs. haloperidol scale: 0–4 improvement in the test score rating after
post-op delirium IV (5 mg); patients the administration of both ondansetron
after on-pump heart were evaluated (from 3.1 to 1.2, percentage improvement
surgery before and 10 min 61.29%, p < 0.01) and haloperidol (from
after the injection 3.1 to 1.3, ± percentage improvement
58.064%, p < 0.01).

Yoon et al., 2013; Patients with Assigned to receive Korean version of Haloperidol, risperidone, olanzapine, and
Observational study; delirium at a tertiary either haloperidol the Delirium Rating quetiapine were equally efficacious and safe
n = 80 level hospital (n = 23), risperidone Scale–Revised-98 in the treatment of delirium. The treatment
(n = 21), olanzapine (DRS-K) response rate was lower in patients over
(n = 18), or 75 years old than in patients under 75 years
quetiapine (n = 18) old, especially for olanzapine.
Maneeton et al., Medically ill patients 25–100 mg/day of DRS-R-98 and total Over the trial period, means (standard
2013; DBRCT, n with delirium quetiapine sleep time deviation) of the DRS-R-98 severity scores
= 52 (n = 24) or were not significantly different between
0.5–2.0 mg/day of the quetiapine and haloperidol groups
haloperidol (n = 28) (–22.9 [6.9] versus –21.7 [6.7]; P = 0.59).
Concluding that low-dose quetiapine and
haloperidol may be equally effective and
safe for controlling delirium symptoms.
Abbreviations: RPCT=randomized placebo control trial; DBRPCT=double blind randomized, placebo controlled trial; OL=open label; trial; DSM=Diagnostic
and Statistical Manual of Mental Disorders; CAM= Confusion Assessment Method; DRS=Delirium Rating Scale; MDAS=Memorial Delirium Assessment Scale;
GI=gastrointestinal; PBO=placebo; Sx=surgery; CPB=cardio-pulmonary by-pass machine; ICU=intensive care unit. APA=antipsychotic agent; OLA=olanzapine;
HAL=haloperidol; RIS=risperidone; AIDS=autoimmune deficiency syndrome; CPM=chlorpromazine; EPS= extra-pyramidal syndrome; LOR=lorazepam; PO=
by mouth.

The second included a head-to-head comparison between
haloperidol and an SGA (risperidone), demonstrating that
delirium scores decreased significantly over the study period
with no significant differences between groups (Han & Kim,
2004). An RPCT 7-day treatment trial compared haloperidol
to olanzapine versus placebo among hospitalized elderly sub-
jects, again showing no difference in the rate of improvement
of delirium with either active agent, but significantly greater
for each than placebo (olanzapine 72.2%; haloperidol 70.4%;
placebo 29.7%; p <0.01, against PBO) (Hu, Deng,  & Yang,
2004). Another study compared enteral haloperidol (avg. dose
6.5 mg/d) versus olanzapine (avg. dose 4.5 mg/d) in the criti-
cal care setting (Skrobik, Bergeron, Dumont,  & Gottfried,
2004). Results show that measures of delirium decreased over
time in both groups, as did the administered dose of as-needed
(PRN) benzodiazepines, with the degree of clinical improve-
ment being similar in both treatment arms.
An RCT comparing the antipsychotics amisulpride versus
quetiapine showed that after intervention, DRS-R-98 scores
were significantly decreased from the baseline in both treat-
ment groups (P <0.001) without group difference, and that
both SGAs were generally well tolerated (Lee, Won, et  al.,
2005). A  study on the use of SGA involved the compari-
son of quetiapine (50 mg twice daily [BID]) versus placebo
in medical ICU patients in three academic centers (Devlin,
Roberts, et  al., 2010). The study found that treatment with
quetiapine was associated with a shorter time to first resolu-
tion of delirium (1.0 vs. 4.5  days; p =.001), a reduced dura-
tion of delirium (36 vs. 120 hrs.; p =.006), less agitation based
on the Sedation-Agitation Scale score (6 vs. 36 hrs.; p =.02),
greater chance to be discharged home versus long-term-care
facility (89% vs. 56%; p =.06), and lower requirement of PRN
haloperidol (3 vs. 4 days; p =.05).
The most recent published blinded RCT compared que-
tiapine (flexible dosing regimen of 25 to 175 mg per day,
in divided doses) versus placebo among medico-surgical
patients (Tahir et al., 2010). Results indicated the quetiapine
group improved more rapidly, 82.7% faster (standard error
[S.E.] 37.1%, P =.026) than the placebo group, in terms of
the DRS-R-98 severity score; and 57.7% faster (S.E. 29.2%,
P =.048) than the placebo group, in terms of the noncogni-
tive subscale.
A Cochrane Database review compared haloperidol with
risperidone, olanzapine, and placebo in the management of
delirium and concluded that there was no significant difference
between low-dose haloperidol (<3.0 mg per day) and the atypi-
cal antipsychotics olanzapine and risperidone (OR 0.63, p =
0.25); that low-dose haloperidol did not have a higher incidence
of adverse effects than the atypical antipsychotics; and that
low-dose haloperidol was effective in decreasing the intensity
and duration of delirium in postoperative patients, compared
with placebo (Lonergan, Britton, Luxenberg, & Wyller, 2007).

878 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
 A subsequent meta-analysis on atypical antipsychotic
agents for the treatment of delirium found that risperidone
was the most thoroughly studied SGA for the management
of delirium (Ozbolt, Paniagua, & Kaiser, 2008). Most studies
found that risperidone was approximately 80%–85% effec-
tive in treating the behavioral disturbances of delirium at
doses of 0.5–4.0 mg per day. On the other hand, olanzapine
was approximately 70%–76% effective in treating the behav-
ioral manifestations of delirium at doses of 2.5–11.6 mg per
day. There were very few studies conducted using quetiapine;
although available data suggest that it also appears to be a safe
and effective alternative to high-potency antipsychotics. In
the limited number of trials comparing SGAs to haloperidol,
haloperidol consistently produced a higher rate (an additional
10%–13%) of extrapyramidal side effects.
Finally, there have been two case reports on the use of
aripiprazole as an effective treatment of delirium (Alao  &
Moskowitz, 2006; Straker, Shapiro, & Muskin, 2006); and a
retrospective case-matched study of aripiprazole vs. haloperi-
dol that found an identical delirium resolution rate (76.2%)
for both agents (Boettger, Friedlander, Breitbart,  & Passik,
2011). In addition, there are two single case reports on the
use of ziprasidone for the management of delirium (Leso &
Schwartz, 2002; Young & Lujan, 2004) and one OL study (n
= 38) on the use of perospirone, a recently developed atypical
antipsychotic with potent serotonin 5-HT2 and dopamine
D2 antagonist activity (Takeuchi et al., 2007).
A systematic literature review of 28 delirium treatment
studies with antipsychotic agents concluded that:
1. Around 75% of delirious patients who receive short-term
treatment with low-dose antipsychotics experience clini-
cal response;
2. This response rate appears quite consistent across differ-
ent patient groups and treatment settings;
3. Evidence does not indicate major differences in response
rates between clinical subtypes of delirium; and
4. There are no significant differences in efficacy for halo-
peridol versus atypical agents (Meagher et al., 2013).
Regarding side-effect incidence, a meta-analysis of all
randomized controlled trials in which SGAs have been com-
pared with conventional drugs concluded that of the SGAs,
only clozapine was associated with significantly fewer extra-
pyramidal symptoms (EPS (p = 0.008) and higher efficacy
than low-potency conventional drugs. A  reduced frequency
of EPS was noted with olanzapine, which was of border-
line significance (p = 0.07) (Leucht, Wahlbeck, Hamann, &
Kissling, 2003). Another study analyzed the data on neu-
roleptics acquired in the German multicenter drug surveil-
lance program database (Projekt “Arzneimittelsicherheit in
der Psychiatrie”) program from 1993 to 2000 and found that
severe adverse drug reactions (ADRs) occurred in 1.1% of the
patients (Bender, Grohmann, et al., 2004). In contrast to the
results from controlled trials, atypical neuroleptics caused
more severe ADRs than did typical neuroleptics. Overall,
SGAs were found to be superior in EPS and urological ADRs,
but when clozapine is excluded, typicals and SGAs have similar
occurrence rates of severe ADRs. Other problems to consider
when choosing an alternative agent include the fact that SGAs
may be associated with weight gain, dyslipidemia, high blood
pressure, and ultimately with cardiovascular disease, diabetes,
and metabolic syndrome (Henderson, 2008). Therefore, when
considering the use of SGAs, clinicians must consider these
factors and weigh potential risks and benefits before prescrib-
ing these agents to a critically ill patient. Finally, there have
been numerous reports attributing the onset of delirium to the
use of SGAs (e.g., clozapine, olanzapine), probably due to their
anticholinergic potential (Bender, Grohmann, et  al., 2004).
But these data are limited to small case reports (see Box 41.7).

Box 41.7 CASE R EPORTS LINKING SGA TO DELIRIUM GENER ATION

• Case reports implicating risperidone use in causing delirium:

• Kato et al. (2005), Psychosomatics
• Morikawa and Kishimoto (2002), Canadian Journal of Psychology
• Tavcar and Dernovsek (1998), Canadian Journal of Psychology
• Chen Chen and Cardasis (1996), American Journal of Psychology
• Case reports implicating quetiapine use in causing the delirium:
• Sim, Brunet, and Conacher (2000), Canadian Journal of Psychology - Balit et al. (2003), Annals of Emergency Medicine
• Miodownik et al. (2008), Clinical Neuropharmacology
• Case reports implicating olanzapine use in causing the delirium:
• Robinson, Burk, & Raman (2003), Journal of Postgraduate Medicine
• Steil (2003) Der Nervenartzt
• Samuels and Fang (2004), Journal of Clinical Psychology - Morita et al. (2004) Journal of Pain & Symptom Management
• Prommer (2005) Journal of Pain & Symptom Management
• Tuglu, Erdogan, & Ebay (2005) Journal of Korean Medicine
• Weizberg et al. (2006) Clinical Toxicology
• Lim, Travino, & Tampi (2006), Annals of Pharmacotherapy

41. D E L I R I U M • 879
 Please note that antipsychotic agents should only be used
short-term for treating agitation in the cognitively impaired
individual. Serious concerns have been raised about the
stroke and mortality risk of atypical antipsychotics when
administered long-term for the management of agitation
in patients with dementia. The dementia antipsychotic
withdrawal trial (DART-AD) found that at 12 months, the
cumulative probability of survival was similar in the active
group and the placebo group (70% versus 77%; p  =  ns)
(Ballard, Hanney, et  al., 2009). Yet, there was a reduction
in survival of the patients who continued to receive antipsy-
chotics (after 12 months) compared with those who received
placebo (24-month survival 46% vs. 71%; 36-month sur-
vival 30% vs. 59%).
Regarding concerns for QTc prolongation, data from the
FDA Division of CardioRenal Drug Products Consultation
suggested that all antipsychotic agents lengthen the QTc to
varying degrees. They based their conclusions on a study of
antipsychotic agents in psychotic subjects with normal base-
line ECGs who were titrated to the “highest tolerated dose” of
one of six antipsychotic agents (see Table 41.14) (FDA, 2000;
Huffman & Stern, 2003). It may be reasonable to assume the
doses used may have been substantially lower than the doses
used for the treatment of hyperactive delirium. We also need
to consider the fact that when these agents are used for the
management of delirium, patients often received parenteral
formulations (when available) due to noncompliance, and
they often experience a number of comorbid medical issues
(Huffman  & Stern, 2003). When antipsychotic agents are
recommended, it is wise to review the patient’s medication
list and identify any other agents with the ability to prolong
QTc. If possible, avoid other medications known to increase
QTc and/or inhibitors of CPY3A4. Before and during the use
of continuous antipsychotic management, you should obtain
a 12-lead ECG, and measure QTc and electrolytes. All elec-
trolyte abnormalities should be corrected (especially K+ and
Mg+) and the QTc pattern monitored. Recommendations
Table 41.14 EFFECTS OF OR ALLY ADMINISTER ED
ANTIPSYCHOTICS ON THE QTC INTERVAL a
MEAN INCR EASE % OF SUBJECTS W ITH
DRUG IN QTC (MS) > 60 MS INCR EASE IN QTC
Thioridazine 35.8 29
Ziprasidone 20.6 21
Quetiapine 14.5 11
Risperidone 10.0 4 occurred after an overdose (LoVecchio, Watts,  & Winchell,
Olanzapine 6.4 4
Haloperidol 4.7 4
Data from Huffman JC, Stern TA. (2003).QTc Prolongation and the use of
antipsychotics: A case discussion. Primary Care Companion to the Journal of
Clinical Psychiatry, 5(6), 278–81.
a
Data adapted from the U.S. Food and Drug Administration’s Center for Drug
Evaluation and Research, Psychopharmacological Drugs Advisory Committee
(FDA, 2000), adapted by Huffman & Stern, 2003.
880 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
are to discontinue antipsychotic use if QTc increases to
>25% of baseline or >500 msec. Of note, studies suggest
that haloperidol may have the lowest ratio of cardiac death
among all dopamine-antagonist agents, both typical and
atypical (Hatta, Takahashi, et  al., 2001; Harrigan, Miceli,
et al., 2004).
Our recommended schedule in the case of hyperactive states
is to use antipsychotic agents (e.g., haloperidol, risperidone, que-
tiapine) at one of the following time schedules: “TID” (thrice
daily) schedule (0600 hrs, 1000, 2000) or “QID” (four times a
day) schedule (0600, 1000, 1600, 2200). The idea is to enhance
behavioral control during the daytime and help promote better
nighttime rest without causing oversedation. Therefore, day-
time doses are significantly lower than the nighttime dose (e.g.,
a 20 mg/d haloperidol dose will be distributed as 5 mg IV q
0600, 1000, and 1600; plus 10 mg at 2200 hrs).
When treating hypoactive delirium, we recommend doses
in the very low range (i.e., haloperidol and risperidone in the
0.25 to 1 mg/24 hr). Data available suggest that the dopamine
excess observed in delirious subjects occurs in all cases, even
hypoactive forms. The same data demonstrate that antipsy-
chotic use helps prevent and treat all forms of deliria, includ-
ing the hypoactive type. In these cases, it is usually given as
a single nighttime dose, just before sundown. Given their
hypoactive state, very sedating agents (e.g., quetiapine, olan-
zapine) are usually not recommended. In the experience of
the author, aripiprazole may prove to be a particularly good
choice for hypoactive cases (usually starting at doses as low
as 1 mg at bedtime), probably due to its partial dopamine
antagonist-agonist properties that may have positive effects
on attention, concentration, and sleep–wake cycle reversal in
delirium, and its minimal muscarinic and histaminic antag-
onist activity (thus minimizing adverse cognitive effects).
Recent reports seem to support this clinical experience, con-
firming the usefulness of aripiprazole, particularly in hypoac-
tive delirium (Alao & Moskowitz, 2006; Straker, Shapiro, &
Muskin, 2006; Boettger & Breitbart, 2011). Aripiprazole had
enjoyed the reputation of being the only SGA not to be associ-
ated with QTc-prolongation or TdP, confirmed by both clini-
cal trials data and clinical experience (Gulisano, Cali, et al.,
2011; Germano, Italiano, et  al., 2014; Li, Luo, et  al., 2014;
Marder et al., 2003). Yet there are at least two case reports of
aripiprazole-induced cardiac arrest due to TdP, when it was
used as a single agent (Nelson & Leung, 2013). There is at least
one case report of aripiprazole-associated QTc prolongation
in a geriatric patient (Hategan  & Bourgeois, 2014). Before
this, all other cases of QTc prolongation had either been asso-
ciated with concomitant use of another antipsychotic agent or
2005; Nelson & Leung, 2013).
Norepinephrine Excess
α 2-Adrenergic Receptors Agonists versus
Conventional Sedative Agents: Effect on Delirium
Prevention
There have been several randomized clinical trials looking at
anesthetic practice and delirium prevention (see Table 41.15).
Table 41.15 Α 2-ADR ENERGIC R ECEPTORS AGONISTS AND ALTER NATIVE SEDATIVE AGENTS VERSUS
CONVENTIONAL ANESTHETICS: EFFECT ON DELIR IUM PR EVENTION AND TR EATMENT

DELIR IUM DELIR IUM INCIDENCE %

STUDY(N = 11) POPULATION INTERVENTION DEFINITION CONTROL/INTERVENTION P-VALUE

Berggren et al., Femoral neck Epidural vs. halothane DSM-III 38% 50% P < 0.05
1987; RCT, n = 57 Fx repair anesthesia (11/29) (14/28)

Williams-Russo B knee Continuous epidural DSM-III 44% 38% p = 0.69

et al., 1992; replacement Sx bupivacaine + fentanyl (11/25) (10/26)
RCT, n = 60 vs. continuous IV
fentanyl

Aizawa GI surgery Usual care vs. BZD DSM-IV 35% 5% p = 0.023

et al., 2002; administration to (7/20) (1/20)
OL, n = 42 promote sleep p-Sx

Maldonado et al., Cardiac Post-op anesthesia DSM-IV Midazolam Propofol Dexmedeto- p <0.001
2003, 2009; valve Sx w/ MID vs. PROP DRS-R-98 50% 50% midine
RCT, n = 118 vs. DEX 3%
(15/30) (15/30)
(1/30)

Pandharipande Med/Surg ICU DEX vs. lorazepam CAM-ICU 3.0d 7.0d p = 0.01
et al., 2007; in mechanical (2 tertiary care centers);
DBRPCT, n = 106 ventilation days alive w/o delirium
or coma

Reade et al., 2009; Tx-agitated ICU IV haloperidol ICDSC 42.2 hrs 20 hrs p = 0.016
randomized, patients 0.5 to 2 mg/hour vs.
open-label, dexmedetomidine Above numbers represent time to
parallel groups 0.2 to 0.7 μg/kg/hr extubation; prolonged intubation
attributed to delirium and agitation
pilot trial; n = 20

Riker et al., 2009;
Med/Surg ICU Midazolam vs. DEX; CAM-ICU 76.6% 54% p<0.001
DBRPCT, n = 375 in mechanical trial conducted in 68 (93/122) (32/244)
ventilation centers in 5 countries

Hudetz Elective PBO vs. IV ketamine Intensive Care 31% 3% p = 0.01

et al., 2009; CABG or valve (0.5 mg/kg) bolus Delirium (9/29) (1/29)
DBRPCT, n = 58 replacement/ during the induction of Screening
repair w CPB anesthesia Checklist
(ICDSC)

Shehabi Elderly patients Morphine vs. DEX CAM-ICU 15% 8.6% p =.088
et al., 2009; after cardiac
DBRPCT, n = 306 surgery

Rubino Acute PBO vs. clonidine Delirium 1.8 ± 0.8 0.6 ± 0.7 p = 0.001
et al., 2010; type-A aortic IV on delirium Detection
DBRPCT, n = 30 dissection repair neurological outcome Score (DDS)→
and respiratory
function
Jakob, Ruokonen, Adult ICU PROP vs. DEX CAM-ICU 29% 18% p = 0.008
et al., 2012; patients (71/247) (45/251)
RDBCT; n = 498 receiving
mechanical
ventilation
Abbreviations: RCT=randomized control trial; RPCT=randomized placebo control trial; DBRPCT=double blind randomized, placebo controlled trial;
OL=open label; trial; DSM=Diagnostic and Statistical Manual of Mental Disorders; CAM= Confusion Assessment Method; DRS=Delirium Rating Scale;
MDAS=Memorial Delirium Assessment Scale; GI=gastrointestinal; PBO=placebo; Sx=surgery; CPB=cardio-pulmonary by-pass machine; ICU=intensive care
unit. APA=antipsychotic agent; BZDP=benzodiazepine; DEX=dexmedetomidine; OLA=olanzapine; HAL=haloperidol; RIS=risperidone; AIDS=autoimmune
deficiency syndrome; CPM=chlorpromazine; EPS= extra-pyramidal syndrome; LOR=lorazepam; PO= by mouth.
The earliest study examined epidural versus halothane anes-
thesia to assess postoperative delirium and found no difference
between the two approaches (Berggren et al., 1987). A second
study compared the effects of continuous epidural bupiva-
caine plus fentanyl versus continuous IV fentanyl anesthesia
on postoperative delirium, and again found no differences
between these approaches (Williams-Russo et al., 1992).
The majority of patients in the ICU, particularly those who
are mechanically ventilated, receive some form of sedation to
reduce their anxiety, encourage sleep, and increase tolerance of
the critical care environment, including multiple IV lines, pain
management, endotracheal tubes, and ventilators. Sedative
and analgesic drugs are among the most commonly prescribed
medications in the ICU (Farina, Levati,  & Tognoni, 1981).
As discussed, sedative agents (mostly GABA-ergic) and opi-
oids may contribute to the development of delirium by one of
six mechanisms: interfering with physiological sleep patterns;
interfering with central cholinergic function; increasing com-
pensatory upregulation of NMDA and kainite receptors and
Ca2+ channels; disrupting the circadian rhythm of melato-
nin release; disrupting thalamic gating function; and leading
to CNS-depressant dependence and withdrawal. Therefore,
sedating alternatives have been thought promising in decreas-
ing medication/sedation-induced delirium.
The first study involving ketamine in delirium preven-
tion was a DBRPCT involving children undergoing dental
repair in sevoflurane-induced anesthesia (Abu-Shahwan  &
Chowdary, 2007). The study demonstrated a substantially
lower incidence of emergence agitation in the ketamine group
compared with placebo (16.6% vs. 34.2%), but no difference
in time to meet recovery room discharge criteria between the
two groups. A second study compared the effects of IV ket-
amine versus placebo on postoperative delirium in patients
undergoing acute type-A aortic dissection repair, which
found a significant reduction in the incidence of delirium in
patients in the ketamine group (3% vs. 31%, p = 0.01) (Hudetz
et al., 2009). Postoperative C-reactive protein concentration
was also lower (p <0.05) in the ketamine-treated patients
compared with the placebo-treated patients.
In an attempt to demonstrate that different sedative strate-
gies may achieve delirium reduction, the author and his team
(Maldonado, 2003) were the first to report on the use of the novel
sedative agent dexmedetomidine (DEX) as an alternative to the
use of benzodiazepines and related agents (e.g., midazolam [MID],
propofol [PRO]) during the postoperative state (Maldonado,
2003; Maldonado et al., 2003, 2004). Post-cardiotomy patients
were selected, given the high incidence of delirium in such
patients (around 57%) nationwide (Ebert, Walzer, Huth,  &
Herrmann, 2001). We studied patients (n = 118)  undergoing
cardiac surgery (i.e., repair or replacement) with cardiopulmo-
nary bypass (Maldonado et al., 2009). Intraoperative anesthesia
for the surgical procedures was standardized for all subjects in all
three studied groups. All procedures were performed via median
sternotomy in conjunction with CPB and induction of moder-
ate hypothermia. After successful weaning from CPB, patients
were started on one of three randomly assigned, postoperative
sedation regimens: DEX (loading dose: 0.4 μg/kg, followed by a
maintenance drip of 0.2–0.7 μg/kg/hour), PRO drip (25–50 μg/
882 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
kg/minute), or MID drip (0.5–2 mg/hour), all titrated to achieve
a target sedation level (i.e., a Ramsay Sedation Score of 3 before
extubation and 2 after extubation). Upon arrival at the ICU, a
standardized protocol for postoperative care was implemented
for all patients. Study results showed there were no significant pre-
operative or intraoperative differences between treatment groups
(e.g., age, sex, American Society of Anesthesiology classification,
bypass time, clamp time, or lowest temperature achieved) (see
Table 41.16). The only real difference in management between
groups was the type of postoperative sedation. Final results dem-
onstrated an incidence of delirium of 3% for patients on DEX,
compared to 50% for PRO and 50% for MID (p <.01). The abso-
lute risk reduction in the incidence of delirium associated with
using DEX was 47% (95% CI 28%–66%) corresponding to a
number-to-treat of 2.1 patients (95% CI 1.5–3.6). Similarly, the
number of delirious days was also significantly lower in the DEX
group compared to PRO and MID (1% vs. 16% vs. 29%, respec-
tively; p <.001).
Since then, two DBRPCT have confirmed the original
findings and demonstrated the delirium-sparing effects of
DEX. The first of these compared DEX with lorazepam (LOR)
in the ICU on mechanical ventilation. The study found that
DEX-treated patients had more delirium-free days (7 days vs.
3  days, p =.01), had a lower prevalence of coma (p <0.001),
and spent more time within sedation goals (p =.04) than
LOR-treated subjects (Pandharipande et al., 2007). Similarly,
a second, much larger trial (i.e., 68 ICUs in five countries)
compared the incidence of delirium among ICU-ventilated
subjects sedated with either DEX or MID and found a lower
delirium incidence (54% vs. 76.6%, p <.001), shorter intuba-
tion time (median time to extubation was 3.7 days DEX vs.
5.6 days MID; p =.01), and less tachycardia and hypertension
in the DEX-treated group (Riker et al., 2009).
A study assessing the effects of different classes of sedative
agents compared variable doses of morphine or dexmedeto-
midine, with open-label propofol titrated to effect (Shehabi
et al., 2009). This study found that the incidence of delirium
was comparable between morphine 22 (15.0%) and DEX 13
(8.6%) (95% CI:  0.256–1.099, p =.088). Yet DEX-managed
patients spent 3 fewer days (2 vs. 5)  delirious (95%
CI:  1.09–6.67, p =.0317), were more likely to be extubated
earlier (95% CI: 1.01–1.60, p =.040), experienced less systolic
hypotension (23% vs. 38.1%, p =.006), and required less nor-
epinephrine (p < 0.001), than morphine-treated patients.
Finally, a subsequent study assessed the role of clonidine IV
(0.5 microg/kg bolus, followed by continuous infusion at 1–2
microg/kg/hr) versus placebo on delirium in post-cardiotomy
patients and found no significant difference in the incidence
of delirium between groups (33% vs. 40%, p = 0.705) (Rubino
et al., 2010). Yet, delirium scores were significantly lower in
patients treated with clonidine compared to placebo (P =
0.001), suggesting that the beneficial effects in delirium pre-
vention previously seen with DEX may extend to all alpha-2
agonists (Rubino et al., 2010).
Despite evidence demonstrating that prophylactic use
of DEX in reducing the incidence of delirium and plenty of
clinical evidence that it is useful in the management of agi-
tation related to hyperactive delirium, there is no study to
Table 41.16 SELECTED POSTOPER ATIVE OUTCOME VAR IABLES FOR CAR DIAC PATIENTS WITH
CAR DIOPULMONARY BYPASS BY INTERVENTION GROUP

DEX MEDE-
TOMIDINE PROPOFOL MIDAZOLA M OVER ALL DEX VS. DEX VS.

(N = 30) (N = 30) (N = 30) P-VALUE PROPOFOL MIDAZOLA M

Delirium

Incidence of delirium (per 1/30 (3%) 15/30 (50%) 15/30 (50%) <0.001 <0.001 <0.001
protocol)

Incidence of delirium 4/40 (10%) 16/36 (44%) 17/40 (44%) <0.001 0.001 0.002

Number of days delirious 2/216 (1%) 45/276 (16%) 75/259 (29%) <0.001 <0.001 <0.001

Average length of delirium days) 2.0 ± 0 3.0 ± 3.1 5.4 ± 6.6 0.82 0.93 0.63

Time Variables

ICU length of stay (days) 1.9 ±.9 3.0 ± 2.0 3.0 ± 3.0 0.11 0,14 0.14

Hospital length of stay (days) 7.1 ± 1.9 8.2 ± 3.8 8.9 ± 4.7 0.39 0.42 0.12

Intubation time (hours) 11.9 ± 4.5 11.1 ± 4.6 12.7 ± 8.5 0.64 0.91 0.34

PRN Medications

Fentanyl (mcg) 320 ± 355 364 ± 320 1088 ± 832 <0.001 0.93 <0.001

Total Morphine equivalents 50.3 ± 38 51.6 ± 36 122.5 ± 84 <0.001 0.99 <0.001

(mg) P‡P

Antiemetic use* 15/30 (50%) 17/30 (57%) 19/30 (63%) 0.58

PRN Medications for the

Management of Delirium**

Lorazepam 1/30 (3%) 7/30 (23%) 6/30 (20%) 0.07 0.06 0.11
Haloperidol 0/30 3/30 (10%) 2/30 (7%) 0.23 0.07 0.15
† Percent of patients who developed delirium.
‡
Sum of average morphine equivalents (fentanyl, oxycodone, and hydrocodone) received in postoperative days 1–3.
* Number of patients who received dolasetron mesylate and/or promethazine HCl in postoperative day 1.
** Average amount over three days. None of these medications were given until a diagnosis of delirium was established.
Source: Maldonado, 2008a.

date confirming its potential in the treatment of delirium.
A randomized, open-label trial of agitated delirium assigned
subjects to receive an infusion of either haloperidol 0.5–2.0
mg/hour or dexmedetomidine 0.2–0.7  μg/kg/hr (Reade
et al., 2009). Results found that DEX significantly shortened
median time to extubation (from 42.5 to 19.9 hours; P =
0.016), significantly decreased ICU length of stay (from 6.5 to
1.5 days; P = 0.004), and, in patients requiring ongoing seda-
tion, DEX cut in half the time propofol was needed (79.5%
vs. 41.2%; P = 0.05). Finally, a recent meta-analysis including
all postoperative prevention strategies found that DEX seda-
tion was associated with less delirium compared to sedation
produced by other drugs (2 RCTs with 415 patients, pooled
RR = 0.39; 95% CI = 0.16–0.95) (Zhang et al., 2013).
The detailed proposed mechanisms for delirium reduction
have been described elsewhere (Maldonado, 2008a). They can
be summarized in six postulated theories: The first suggests
that DEX has an intrinsic delirium-sparing effects, which
may be explained by DEX’s pharmacological characteristics.
The second theory suggests that the reason patients had sig-
nificantly less delirium in the DEX group was not because of
its use per se, but because those patients were not exposed to
other sedative agents with much greater delirium potential
(e.g., GABA-ergic agents, such as propofol and midazolam,
and lower doses of opioid agents). Third, data suggest that
the neuroprotective effect of DEX is mediated by activation
of the alpha2A adrenergic receptor subtype (Ma et al., 2004).
Fourth, it is worth noticing that DEX has been shown to have
significant neuroprotective effect, in vivo and in vitro, against
glutamate-induced damage via an increased astrocyte expres-
sion of BDNF through an extracellular signal-regulated
kinase-dependent pathway (Engelhard et al., 2003; Ma et
al., 2004; Sato, Kimura, Nishikawa, Tobe, & Masaki, 2010;
Degos et al., 2013). Fifth, it is also hypothesized that DEX

41. D E L I R I U M • 883
increases the expression of active (autophosphorylated) focal
adhesion kinase, a non-receptor tyrosine kinase playing a piv-
otal role in cellular plasticity and survival (Dahmani, Rouelle,
Gressens, & Mantz, 2005). The sixth theory involves DEX’s
unique mechanism for inducing sleep. Data suggest that DEX
induces a qualitatively similar pattern of c-Fos expression as
seen during normal NREM sleep by acting on endogenous
sleep pathways (i.e., a decrease in the locus ceruleus and tuber-
omammillary nucleus and an increase in the ventrolateral pre-
optic nucleus) (Nelson et al., 2003).
It is certainly possible that a combination of any or all of
these six mechanisms work together, leading to the results
obtained in the above-mentioned studies. It is this author’s
experience that all centrally acting α2-adrenergic recep-
tors agonists are to some extent useful in the management
of hyperactive delirium. In fact, our team uses clonidine to
help transition patients off IV-DEX with good success (see
Table 41.17). We also use clonidine, orally and transdermally,
for prevention of alcohol withdrawal, with excellent results.
Another α2-adrenergic receptor agonist worth considering
is guanfacine. This agent is even more α2/ α1 selective and is
available in oral dose form. This makes it an excellent alterna-
tive to DEX (which requires an IV drip and a monitored bed)
and is less hypotensive than clonidine. Our team has been
using this agent, as we have clonidine, both as adjunct treat-
ment of hyperactive delirium and as a primary agent in the
prevention and treatment of alcohol withdrawal.
Glutamate Excess
Glutamate Antagonists and Ca+ Channel
Modulators for Delirium Management
A number of agents with antiglutamatergic and Ca+ chan-
nel blocking qualities are worth considering here:  lamotrig-
ine, amantadine, memantine, gabapentin, and valproic acid
(VPA) (see Table 41.18).
Both amantadine and memantine have been recognized as
having neuroprotective effects, probably mediated by their pro-
tection from glutamate-induced exocytosis (by blocking exces-
sive N-methyl-D-aspartic acid receptors without disrupting
Table 41.17 CENTR ALLY ACTING Α 2-ADR ENERGIC R ECEPTOR AGONISTS
DRUG Α 2/Α 1SELECTIVITY DT ½
Guanfacine 2,640 2.5 hr
Dexmedetomidine 1,600 6m
Medetomidine 1,200
Clonidine 220 11 m
Methyldopa 12 m
Guanabenz 60 m
Abbreviations: dT1/2= time to onset; eT1/2= elimination half-life; PO=by mouth; IV=intravenous; TDS= transdermal delivery (patch).
884 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
physiological synaptic activity, thus preventing excessive cal-
cium influx into neurons—believed to be the key early step in
GLU-induced exocytosis); reducing the release of proinflam-
matory factors from activated microglia; inducing expression
of neurotrophic factors, such as glial cell line–derived neuro-
trophic factor in astroglia; and limiting oxidative injury and den-
dritic degeneration induced by anticholinesterase neurotoxicity
(Giacino & Whyte, 2003; Zaja-Milatovic, Gupta, Aschner, &
Milatovic, 2009; Ossola et al., 2011; Kutzing, Luo, & Firestein,
2012). Thus, it makes sense to consider their use in various syn-
dromes associated with excess glutamate and subsequent cogni-
tive decline (e.g., traumatic brain injury, stroke, delirium). In
fact, studies have demonstrated that memantine may be effec-
tive in reducing the damage induced by acute ischemia/reperfu-
sion (Yigit et al., 2011), while amantadine has been shown to
enhance cognitive recovery and minimize delirium after severe
TBI in humans (Giacino et al., 2012).
A DBPCRT compared the use of gabapentin as an add-on
agent in the treatment of postoperative pain (i.e., spine sur-
gery under general anesthesia) to reduce the occurrence of
postoperative delirium (Leung et al., 2006). Subjects received
either placebo or gabapentin 900 mg administered orally one
to two hours before surgery, and continued once a day for the
first 3 postoperative days. Results demonstrated that gabapen-
tin was superior to placebo in reducing delirium occurrence
(0% vs. 42%, p = 0.045). Even though the study authors attrib-
uted this to gabapentin’s opioid-sparing effect, they failed to
consider gabapentin’s true modes of action (i.e., modulation
of voltage-sensitive Ca2+ channels, NMDA receptor antag-
onism, activation of spinal alpha-2 receptors, attenuation of
Na+ dependent action potentials) as potential mediators of
its deliriolytic effect.
Even though many of us routinely use VPA (either oral
or IV) in the management of agitated delirious patients who
either are not responsive or cannot tolerate conventional
treatment, there is very little literature investigating the effec-
tiveness of this practice. In fact, there are no RCT or even OL
studies on the matter. The only published report on the effec-
tiveness of VPA for delirium treatment comes from a report of
six cases (Bourgeois, Koike, Simmons, Telles, & Egglestone,
PRODUCT PROTEIN
ET ½ AVAILABILITY BIOAVAILABILITY BINDING
17 hr PO ~100% 70%
2 hr IV 70–80% 94%
13 hr PO 100% PO 40%
TDS 60% TDS
105 m PO/IV 50% <20%
6 hr PO 75% 90%
Table 41.18 GLUTAMATE ANTAGONISTS AND CA+ CHANNEL MODULATORS FOR DELIR IUM TR EATMENT

PRODUCT BIO- PROTEIN

DRUG T½ AVAILABILITY AVAILABILITY METABOLISM BINDING MECHANISM ACTION

Lamotrigine 25 hr PO ~100% Hepatic 55% • Stabilizes neuronal membranes

• Inhibits voltage-sensitive Na+
channels and/or Ca+ channels→ ↓
cortical GLU release
• Calcium channel blockers
• Excitatory amino acid antagonists

Amantadine 17 ± 4 hr PO 86–90% None 67% • NMDA receptor antagonist

Renal excretion • ↑ synthesis and release of
dopamine

Memantine 60–80 hr PO 100% Mostly 45% • Noncompetitive NMDA receptor

unchanged renal antagonist
excretion • Blocks the effects of excessive
levels of GLU
• Some Ca+ channel blockade
• 5HT3 antagonist

Gabapentin 5–7 hr PO 60% None <3% • Voltage-gated Ca+ channel

Renal excretion blockade → ↓ cortical GLU release
• NMDA antagonism
• Activation of spinal
alpha2-adrenergic receptors
• Attenuation of Na+ dependent
action potential
VPA 9–16 hr PO/IV 90% Hepatic 90% • GABA transaminase inhibitor
conjugation → ↑ GABA
• Inhibits voltage-sensitive Na+
channels→ ↓ cortical GLU release
• ↓ release of the
epileptogenic amino acid
gamma-hydroxybutyric acid
(GHB)
Abbreviations: T1/2=elimination half-life.

2005). The authors reported that “in all six cases, the use of
VPA combined with conventional antidelirium medications
resulted in improved control of behavioral symptoms without
significant side effects.” In all cases, standard psychotropic
medication regimens commonly used for the management of
delirium and/or nonspecific agitation in the general hospital
setting (i.e., antipsychotics and/or benzodiazepines) had ini-
tially been tried with limited success, and in all cases the addi-
tion of VPA to the existing regimens resulted in improved
clinical response. Daily doses ranged between 0.5 to 2.5 grams
in divided doses, without significant side effects and with dis-
cernible therapeutic benefits.
Similarly, the Stanford’s Psychosomatic Medicine
Research Lab presented a larger case series (n = 16) equally
demonstrating the usefulness and safety of VPA in the man-
agement of agitated/hyperactive delirium, as an add-on to
ineffective response to conventional therapy (Sher, Lolak,
Miller, & Maldonado, 2013). In this series, the dose ranged
between 0.5 to 3.0 grams per day in divided doses. All subjects
in this case series improved in their clinical status with even-
tual resolution of delirium. In the case of intubated patients,
addition of VPA frequently allowed prompt extubation, since
agitation and related desaturation were better controlled. No
significant side effects were identified in any of the subjects,
with the exception of nonclinically significant decreases in the
number of platelets in few subjects. Of note, as in the case of
SGAs, there have been case reports on VPA-induced delirium.
The availability of IV (Depacon) and elixir (Depakene, ideal
for administration via feeding tube, if needed) forms allow
administration in non-cooperative subjects, or those intu-
bated and unable to take PO medication. The IV formulation
even allows for loading to quickly establish therapeutic serum
levels, then continuing for maintenance or converting to PO
when medically indicated. Others have described the useful-
ness of VPA in the management of destructive and aggressive
behavior associated with brain injury (Wroblewski, Joseph,
Kupfer,  & Kalliel, 1997)  and dementia (Haas, Vincent,
Holt, & Lippmann, 1997; Narayan & Nelson, 1997; Tariot,
1999; Porsteinsson, Tariot, et al., 2003; Sival, Duivenvoorden,
et al., 2004). As in any patient receiving VPA, close monitor-
ing of liver function tests, bilirubin, platelet count, and amy-
lase are advisable, especially in the critically ill patient.

41. D E L I R I U M • 885
 Cholinergic Deficit
Acetylcholinesterase Inhibitors for Delirium
Prevention
Despite the logical premise behind the prophylactic use of
acetylcholinesterase inhibitor agents, studies have not con-
sistently demonstrated positive results (see Table 41.19). The
first publication suggesting the promise of acetylcholinester-
ase inhibitors in delirium prevention comes from a study on
the use of rivastigmine for delirium prevention among hos-
pitalized elderly patients who had been chronically treated
with rivastigmine for dementia (Dautzenberg, Mulder, Olde
Rikkert, Wouters,  & Loonen, 2004). A  retrospective chart
review showed that those on rivastigmine had a significantly
lower incidence of delirium compared to subjects not treated
(45.5% vs. 88.9%; p <0.05). Similarly, a prospective study of
elderly subjects suffering from vascular dementia randomly
placed patients on either rivastigmine (6 mg/d) or aspirin (100
mg/d) and followed them at regular intervals for 24 months
(Moretti, Torre, Antonello, Cattaruzza,  & Cazzato, 2004).
The results suggested that those on rivastigmine had a sub-
stantially lower incidence of delirium compared to aspirin
(40% vs. 62%; p <0.001). Also of significance, when sub-
jects developed delirium, the mean duration of delirium was
shorter (mean duration 4  ± 1.71  days vs. 7.86  ± 2.73  days;
p <0.01), and the use of benzodiazepine and neuroleptic
agents for management of agitation was significantly less
in the rivastigmine group (p <0.05). Furthermore, analysis
of behavioral data (as measured by Behavioral Pathology in
Alzheimer’s Disease (BEHAVE-AD) scale), found that total
scores in the rivastigmine group were significantly improved
over baseline and at the end of the study (all p <0.001). Further
sub-analysis of the BEHAVE-AD individual items indicated
that rivastigmine provided benefits on all items of the scale,
except for delusions, throughout the study.
Unfortunately, the only DBRCT of rivastigmine failed
to demonstrate efficacy in the prevention of postoperative
delirium. In a study in cardiac surgery patients comparing
rivastigmine (1.5 mg/d for 3 days preoperatively) against pla-
cebo found no benefits of the active drug use (Gamberini
et al., 2009).
Similarly, results have been negative with the use of
donepezil. The first study was a DBRPCT involving elderly
patients undergoing elective total joint replacement surgery

Table 41.19 ACETYLCHOLINESTER ASE INHIBITORS IN DELIR IUM PR EVENTION

DELIR IUM INCIDENCE %

DELIR IUM
STUDY(N = 7) POPULATION INTERVENTION DEFINITION CONTROL INTERVENTION P-VALUE

Dautzenberg (P) ≥65y/o Patients who Retrospective 88.9 45.5 P < 0.05
et al., 2004 hospitalized used rivastigmine chart review of (26/29) (4/11)
OL, Retrospective demented chronically with a geriatric service
review, n = 51 patients randomly selected consultations
subgroup of all patients
not treated

Moretti et al., 2004 (P) ≥65y/o–o/p, Cardioaspirin vs. CAM 62 40 P < 0.001

RCT, n = 230 w/ vasc dementia rivastigmine PO qD BEHAVE-AD (71/115) (46/115)
(24-month f/u)

Liptzin et al., 2005 (P) Elderly PBO vs. donepezil DSM-IV 17.1 20.5 p = 0.69
DBRPCT n = 80 elective total joint (14 days pre + 14 days (7/41) (8/39)
replacement post Sx)

Sampson et al., 2007 (P) Elderly PBO vs. donepezil 5 DSI 35.7 9.5 p = 0.08
DBRPCT, n = 33 elective hip mg immediately p-Sx (5/14) (2/19)
replacement + 3 days

Oldenbeuving (P) Delirium p Rivastigmine 3→ DRS ≥ 12 In 16/17 (94%) delirium severity

et al., 2008 CVA 12 mg/d; no PBO improved; mean decrease 14.8→8.5; mean
N = 26 duration: 6.7 days; no side effects

Gamberini (P) Cardiac Sx PBO vs. PO CAM 30 32 p = 0.8

et al., 2009 under CPB rivastigmine 1.51 (17/57) (18/56)
DBRPCT, n = 120 pre-op, until POD #6

van Eijk et al., (Tx) >18 y/o in 2 arms, both receiving CAM-ICU 3d 5d p = 0.06
2010DBRPCT, ICU haloperidol, trial PBO
n = 109 vs. rivastigmine Above refers to delirium duration
in days; trial halted due to mortality
in rivastigmine (n = 12, 22%) was >
than in placebo group (n = 4, 8%;
p = 0.07)
1
Rivastigmine-treated patients who experienced delirium had a shorter duration, lower use of benzodiazepine and neuroleptic for management of agitation, and
improvement in all behavioral aspects measured by the BEHAVE-AD.

886 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
(Liptzin, Laki, Garb, Fingeroth, & Krushell, 2005). Subjects
in the active treatment group received donepezil or placebo
for 14 days before surgery and 14 days afterward, and no sig-
nificant differences between groups were found (p = 0.69).
A second RCT also failed to demonstrate efficacy of donepezil
(5 mg immediately postoperatively and every 24 hrs thereafter
for the first 3 postoperative days) over placebo in preventing
postoperative delirium after elective total hip replacement
surgery in older people without preexisting dementia (9.5%
donepezil vs. 35.7% PBO; p = 0.08) (Sampson et al., 2007).
Even though this was a negative study, the trend was in the
right direction (lower incidence of delirium; shorter length of
stay [9.9 vs. 12.1 days; p = 0.09] and shorter delirium duration
[1.5 vs. 1.8 days; p = 0.83]), and the study had Type-II errors
(i.e., it was grossly underpowered).
The results of the acetylcholinesterase trials seem to sug-
gest that these agents are either (a)  just not good for delir-
ium prevention; or (b) they need to be used for much longer
periods of time (as in the case of Dautzenberg and Moretti’s
study); or (c)  they need to be used at doses much higher
than we currently use them in order to achieve acute clinical
efficacy.
A note of warning: more recently a DBRPCT compar-
ing rivastigmine to placebo was stopped after members of
the data safety and monitoring board unblinded the results
and performed an interim analysis, halting the study due to
concerns about a higher mortality rate in the rivastigmine
group (p = 0.07) (van Eijk et al., 2010). The authors failed
to explain the mechanism by which rivastigmine use may
have led to the higher mortality reported. It is difficult to
interpret their report, as less than 25% of the sample had
been studied, and the causes of death for those on the study
drug were heterogeneous. It is clear that the study relied on
a very fast rivastigmine titration (essentially doubling the
dose every 3 days until reaching the goal of 6 mg BID by day
10). This could have contributed to untoward side effects,
which could have potentially worsened underlying cardio-
vascular or metabolic processes. Unlike in other delirium
studies, the population in this study was not homoge-
neous, as it involved patients with all types of disease in the
ICU. As the authors reported, “we cannot exclude that the
recorded difference in mortality is due to chance” (van Eijk
et al., 2010, p.1835).
Acetylcholinesterase Inhibitors for Delirium Treatment
Addressing the theory that proposes that delirium is caused
by a central cholinergic deficiency state, some researchers
and clinicians have experimented with the use of acetylcho-
linesterase inhibitor agents. Most of the published data con-
sist of small series of case reports associated with the use of
rivastigmine in the treatment of delirium in older persons
(Dautzenberg, Mulder, et al., 2004; van den Bliek & Maas,
2004)(van den Bliek and Maas 2004). There have been at
least 19 papers, mostly case reports, suggesting that acetyl-
cholinesterase inhibitor agents (e.g., donepezil, galantamine,
physostigmine, rivastigmine) may be effective in the treat-
ment of delirium (see Box 41.8).
41. D E L I R I U M
Box 41.8 CASE R EPORTS SUGGESTING A POSITIVE
EFFECT OF ACETYLCHOLINESTER ASE INHIBI-
TORS IN THE TR EATMENT OF DELIRIUM
• Burt, 2000
• Bruera, Strasser, et al., 2003
• Dautzenberg, et al., 2003, 2004
• Fisher et al., 2001
• Gleason, 2003
• Hasse & Rundshagen, 2007
• Hori, Tominaga, et al., 2003
• Kaufer, Catt, et al., 1998
• Kobayashi et al., 2004
• Logan & Stewart, 2007
• Moretti et al., 2004
• Palmer, 2004
• Rabinowitz, 2002
• Weizberg et al., 2006
• Wengel, Roccaforte, & Burke 1998
• Wengel, Burke, & Roccaforte, 1999
Physostigmine and Delirium
Acetylcholine deficiency has been postulated as one of the
potential causes of delirium; whether this is caused by nor-
mal physiological causes (e.g., aging) or due to exogenous fac-
tors (e.g., use of anticholinergic substances). Thus, among the
tools to treat delirium, particularly when it is presumed to be
due to the ingestion of substances with anticholinergic poten-
tial, we should consider the potential use of physostigmine.
Physostigmine, a short-acting acetylcholinesterase inhibi-
tor, increases synaptic acetylcholine concentrations and can
overcome the postsynaptic muscarinic receptor-blockade pro-
duced by anticholinergic agents. As a tertiary amine, it can
pass freely into the CNS and reverse both central and periph-
eral anticholinergic effects.
Many reports have demonstrated its utility and safety in
cases when delirium has been caused by medication overdose
(whether accidental or intentional) (Stern, 1983); Lipowski,
1992; Beaver & Gavin, 1998; Richardson, Williams, & Carstairs,
2004; Eyer et al., 2011; Hail, Obafemi, & Kleinschmidt, 2013).
Physostigmine has been successfully used to treat emergence
delirium in both adults (Brown, Heller,  & Barkin, 2004;
Haase  & Rundshagen, 2007)  and pediatric patients (Funk,
Hollnberger, & Geroldinger, 2008). It has been reported that
physostigmine attenuates several withdrawal states, especially
alcohol delirium and opiate and nitrous oxide withdrawal
syndromes; and it may offer a protective mechanism against
hypoxic damage of the brain (Powers, Decoskey, & Kahrilas,
1981; Rupreht, Schneck, & Dworacek, 1989).
• 887
 Despite many psychiatrists’ concerns about physostig-
mine administration, its use as a diagnostic test appears to be
rather safe. In a series of 39 adults treated with varying doses
of physostigmine (range 0.5–2.0 mg), 56% of all cases experi-
enced full reversal of delirium. Among the cases known to be
antimuscarinic in etiology (e.g., overdose with a known sub-
stance) 100% of patients experience full reversal. In this sam-
ple, no patient experience dysrhythmias or signs of cholinergic
excess, or required the administration of atropine. One (2.6%)
in 39 patients experienced a brief convulsion, with no known
adverse sequelae (Schneir et al., 2003). Similarly, others have
reported that among 52 consecutive patients with suspected
anticholinergic delirium, physostigmine controlled agitation
and reversed delirium in 96% and 87% of cases, respectively
(Burns, Linden Graudins, Brown, & Fletcher, 2000). Again,
no significant side effects were reported.
Given its safety profile and effectiveness, physostigmine
should be considered when a delirious patient’s examination
exhibits signs of a central anticholinergic state (e.g., confu-
sion, sinus tachycardia, markedly dilated and fixed pupils, dry
mouth, hypoactive bowels sounds, dry and flushed skin; see Box
41.9) and/or when it is known that the patient’s altered mental
status is due to the use of known anticholinergic substances
(e.g., diphenhydramine). An initial physostigmine dose of 1–2
mg (0.5 mg in children) given intravenously over 3 to 5 minutes
is the recommended dose. If the response is incomplete, addi-
tional doses of 0.5–1.0 mg every 5 minutes may be given until
delirium resolves or there are signs of cholinergic excess (e.g.,
diaphoresis, salivation, vomiting, diarrhea). A  prolonged PR
interval (>200 ms) or QRS complex (>100 ms and not related
to bundle branch block) interval on ECG are considered the
only absolute contraindications for physostigmine use.
Melatonin and Sleep Deficit
Sleep deprivation is one of the major theories on the develop-
ment of delirium, whether caused by medication effect, envi-
ronmental factors, or patient characteristics. Many sedative
and hypnotic agents may worsen sleep and thus are not recom-
mended. As described in the prevention section, various seda-
tive hypnotic agents are commonly used in the medical setting,
but most of them have significant problems, including caus-
ing or worsening delirium, cognitive clouding, and respiratory
depression.
Box 41.9 ANTICHOLINERGIC SIDE EFFECTS
Cognitive impairment
Delirium
Dilated and fixed pupils, blurred vision
Dry and flushed skin
Dry mouth
Fever
Hallucinations
Hypoactive bowels sounds, constipation, ileus
Sinus tachycardia
Urinary retention
888 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
Various sedative hypnotic agents are commonly used in
the medical setting. Nevertheless, benzodiazepines and the
benzodiazepine-receptor agonists (e.g., zolpidem) may cause or
exacerbate delirium. Other sedating agents, such as antidepres-
sants (e.g., trazodone, mirtazapine) and antipsychotics (e.g., olan-
zapine) have been used off-label to promote sleep, but these agents
have not been tested in the critically ill population. Clinicians
must also consider factors such as drug–drug interaction and
medication half-lives when prescribing. For example, mirtazapine
and trazodone may indeed promote night sleep, but their effects
may last well into the next day, interfering with cognition, atten-
tion, and concentration. Sedative agents with high anticholiner-
gic load, such as antihistaminic agents (e.g., diphenhydramine,
hydroxyzine) or tricyclic antidepressants (e.g., amitriptyline)
should be avoided, as they will aggravate delirium even if they are
immediately effective in promoting sleep. Similarly, benzodiaz-
epines should also be avoided if at all possible.
An alternative to these is the use of non-benzodiazepine
agents, such as melatonin or melatonin agonists (i.e., ramelt-
eon). Melatonin has been shown to play an important role
in the regulation of circadian rhythm and maintenance of a
physiological, well-regulated sleep-wake pattern (Brzezinski,
1997). Multiple studies have demonstrated a number of
abnormalities linking melatonin and delirium, includ-
ing (a)  irregular patterns of melatonin circadian rhythm
(Miyazaki et  al., 2003; Olofsson, Alling, Lundberg,  &
Malmros, 2004); and (b)  decreased melatonin secretion
(Shigeta et al., 2001; Guo, Kuzumi, Charman, & Vuylsteke,
2002)  and abnormalities in the 24-hour urinary excretion
of 6-sulphatoxymelatonin (6-SMT); the chief metabolite of
melatonin (Balan, Leibovitz, et al., 2003). (See Table 41.20
for publications on melatonin abnormalities in delirium.) It
is also theorized that melatonin may have protective effects,
at least in cases of global cerebral ischemia/reperfusion
injury, which may have implications for delirium onset (Sun,
Lin, et  al., 2002; Zhang, Guo, et  al., 2002; Cheung, 2003;
Cervantes, Morali, et  al., 2008). In fact, melatonin may
have a number of beneficial physiological effects which may
prove benefical in the management of medically ill, delirious
patients (see Box 41.10).
Melatonin in Delirium Prevention
Several case reports have described the successful use of melato-
nin in preventing postoperative delirium (Hanania & Kitain,
2002). A small DBRPCT demonstrated that the use of mela-
tonin (10 mg q HS) was associated with longer and better sleep
quality (p = 0.04) (Bourne et al., 2008). A recent DBRPCT
(n  = 222)  compared the effects of melatonin versus placebo
among elderly patients undergoing hip arthroplasty under
spinal anesthesia and found that those receiving melatonin
showed a statistically significant decrease in the percent-
age of postoperative delirium (9.43% vs. 2.65%, p = 0.003)
(Sultan, 2010). Similarly, a recent DBRPCT (n = 145) among
tertiary-care medicine service elderly individuals demon-
strated that the administration of low-dose (i.e., 0.5 mg) mela-
tonin was associated with a lower risk of delirium (12.0% vs.
31.0%, p = 0.014), adjusted for dementia and other comorbidi-
ties (Al-Aama et al., 2011).
Table 41.20 DELIR IUM-ASSOCIATED MELATONIN ABNOR MALITIES

DELIR IUM
STUDY POPULATION INTERVENTION DEFINITION R ESULTS

Shigeta Men and women None; observational CAM All patients without delirium showed nearly identical
et al., 2001 undergoing preoperative and postoperative melatonin secretion for
N = 29 laparotomy for 24 hours.
digestive disease Conclusions: Abnormal melatonin secretion may be
involved in postoperative sleep disturbances, which
triggered delirium in elderly patients.

Guo et al., Males undergoing None; observational Only three patients regained circadian secretion of cortisol.
2002N = 12 elective CABG with Conclusion: melatonin and cortisol secretion were
CPB disrupted during cardiac surgery with CPB and in the
immediate postoperative period

Miyazaki s/p esophagectomy None; observational CE A significant correlation was seen between ICU psychosis
et al., 2003 for CA and an irregular melatonin circadian rhythm (P =.0001).
N = 41

Balan Patients who None; patients ICD-10 24-hr urinary excretion of 6-sulphatoxymelatonin
et al., 2003 developed were divided DRS (6-SMT)—the chief metabolite of melatonin.
N = 31 delirium during into: hyperactive (7), Among hyperactive patients, the levels of 6-SMT were
hospitalization hypoactive (10), mixed lower during the acute delirium than after recovery
(14) (p<0.001). Among hypoactive patients, 6-SMT levels
were higher during the acute delirium than after
recovery (p < 0.01). No difference in mixed.
Yoshitaka, Egi, 40 postoperative None; observational CAM There was no difference in preoperative melatonin
et al., 2013 patients in intensive concentration. The Δ melatonin concentration at
N = 40 care unit 1-hr post-op was significantly lower in patients with
delirium than in those without delirium (P =.036).
After adjustment of relevant confounders, Δ melatonin
concentration was independently associated with risk of
delirium (OR, 0.50; P =.047).

Melatonin and Melatonin Agonists
in Delirium Treatment
Several case reports have described the successful use of mela-
tonin in treating severe postoperative delirium unresponsive
to antipsychotics or benzodiazepines (Hanania  & Kitain,
2002). As part of the Sultan (2010) DBRPCT noted above
(n = 222), comparing the effects of melatonin versus placebo
among elderly patients undergoing hip arthroplasty under spi-
nal anesthesia, a number of subjects developed postoperative
delirium and received postoperative melatonin for treatment
(n = 62). It was found that 58% of subjects experienced delir-
ium resolution after treatment administration (Sultan, 2010).

Box 41.10 MELATONIN PHYSIOLOGICAL EFFECTS

• Melatonin plays important roles in multiple bodily functions, which may have implications for the development of delirium in the
medically ill:
• Chronobiotic effect (affecting aspects of biological time structure)
• Sleep–wake cycle regulatory effects
• Helps reset circadian rhythm disturbances
• Extensive antioxidant activity (with a particular role in the protection of nuclear and mitochondrial DNA)
• Extensive antiinflammatory activity
• Anti-nociceptive and analgesic effects
• Melatonin receptors appear to be important in mechanisms of learning and memory
• Inhibits the aggregation of the amyloid beta protein into neurotoxic microaggregates responsible for the neurofibrillary tangles
characteristics of Alzheimer’s disease, and it prevents the hyperphosphorylation of the tau protein.
Adapted from Maldonado (Maldonado, 2008a).
Darlington, L. G., C. M. Forrest, G. M. Mackay, R. A. Smith, A. J. Smith, N. Stoy and T. W. Stone (2010). “On the Biological Importance of the 3-hydroxyanthranilic
Acid: Anthranilic Acid Ratio.” Int J Tryptophan Res 3: 51–59.

41. D E L I R I U M • 889
 Of note, a systematic literature search of all papers pub-
lished on the use of melatonin for the treatment in dementia
revealed a significant improvement in sundowning/agitated
behavior (de Jonghe, Korevaar, van Munster,  & de Rooij,
2010). This may be particularly important given the presence
of circadian rhythm disturbances, like sundowning, in both
dementia and delirium. Similarly, case reports have suggested
the usefulness of ramelteon, a melatonin agonist, in the treat-
ment of delirium (Kimura et al., 2011).
Similarly, there are two case reports of the successful use
of ramelteon (a novel selective melatonin-receptor agonist) in
the treatment of patients with delirium (Kimura et al., 2011;
Furuya et al., 2012). (See Table 41.21 for a summary of pub-
lished case reports and studies on the use of melatonin for the
treatment of delirium.)
When it comes to sleep and delirium, the clinician
must consider both risks and benefits of the intervention.
In general, lack of sleep may lead to delirium, especially in
patients at high risk (e.g., the elderly; the seriously medi-
cally ill); thus, promoting sleep is paramount. The first step,
as always, is prevention. Therefore, the implementation of
nonpharmacological sleep protocols as described above is
important. If the patient still has difficulty sleeping, the use
of melatonin (or melatonin agonists) is probably the best
first pharmacological option. When this is not enough, clini-
cians should consider the use of trazodone, followed by mir-
tazapine. If needed, zolpidem may be an acceptable choice.
Compared to conventional benzodiazepine agents, zolpidem
may have a lower deliriogenic potential. Certainly, some of
the other delirium treatment alternatives already discussed
above may be beneficial for sleep promotion, such as the use
of nighttime gabapentin or VPA or the use of low-dose que-
tiapine. As a general rule, it is advisable to avoid the use of
diphenhydramine due to its anticholinergic effects. Although

Table 41.21 MELATONIN AND ANALOG FOR DELIR IUM MANAGEMENT

DELIR IUM
STUDYN = 7 POPULATION INTERVENTION DEFINITION R ESULTS

Bourne et al., 2008 S/p tracheostomy to Melatonin (MEL) Bispectral Melatonin use was associated with a 1-hour increase
N = 24 assist weaning from 10 mg PO @ 2000 index (BIS) in nocturnal sleep (P = 0.09) and a decrease in BIS
vent AUC indicating “better” sleep.
DBPCT
Melatonin use was associated with increased
nocturnal sleep efficiency.

Al-Aama, Brymer, Individuals aged ≥65 Randomized to MEL CAM Melatonin was associated with a lower risk of
et al., 2011 y/o admitted through 0.5 mg vs. PBO every delirium (12.0% vs. 31.0%, p = 0.014).
N = 145 the ED to a medical night for 14 days or
unit until discharge.

Sultan, 2010 ≥ 65 y/o scheduled Randomized to: Abbreviated The melatonin group showed a statistically
N = 300 for hip arthroplasty 1. PBO Mental Test significant decrease in the percentage of
under spinal (AMT) postoperative delirium to 9.43%.
anesthesia 2. Melatonin 5 mg
POD: PBO—32.7 %; MEL—9.4 % (p = 0.003);
3. Midazolam 7.5 mg MID—44 and (p = 0.245)
4. Clonidine 100 μg CLO—37.3 % (p = 0.629); Melatonin was
successful in treating 58.06% of patients who
suffered postoperative delirium.

de Jonghe Meta-analysis Various Nine papers, including RCTs (n = 243), and five case
et al., 2010 series (n = 87) were reviewed. Two of the RCTs found
Review a significant improvement on sundowning/agitated
behavior. All five case series found an improvement.

de Jonghe, van ≥65 y/o admitted for Randomized to: CAM Ongoing

Munster, et al., 2011 surgical repair of hip PBO
N = 452 fracture
Melatonin 3 mg @
2100

Kimura et al., 2011 Pts >59 y/o, medically Open label; ramelteon DSM-IV-TR All three cases demonstrated significant
N = 3 (case report) ill 8 mg q HS MDAS-Jap improvement in delirium scores as measured by
MDAS, with steady improvement over 7 days,
ramelteon 8 mg at HS.
Furuya et al., 2012 Mostly demented, Open label; ramelteon DSM-IV-TR Authors reported the successful treatment of five cases
N = 5 (case report) medically ill patients 8 mg of delirium within 1 day, after ramelteon 8 mg at HS.

Abbreviations: MEL=melatonin; BIS= Bispectral index (BIS) monitor.

890 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
low-dose doxepin is an excellent alternative for many adult
patients, it is unclear if it would still have enough anticholin-
ergic effect to make it wise avoid it. Certainly, the use of con-
ventional benzodiazepine agents is discouraged due to their
inherit deliriogenic effect.
S L E E P R E S TOR AT ION I N DE L I R I U M
T R E AT M E N T
As described above, in delirium treatment we highly rec-
ommend starting with the implementation of nonpharma-
cological sleep protocols. If the patient still has difficulty
sleeping, the use of melatonin (or melatonin agonists) is
probably the best first pharmacological option (e.g., 3 mg
PO q, 2000; for prophylaxis). Several case reports have
described the successful use of melatonin (e.g., 1–36 mg HS
PO, for treatment) and melatonin agonists in treating severe
postoperative delirium that is unresponsive to antipsychot-
ics or benzodiazepines (Sultan, 2010; Kimura et  al., 2011;
Furuya et al., 2012).
If that fails, clinicians should consider the use of
non-benzodiazepine agents, such as trazodone or mirtazap-
ine. If absolutely necessary, zolpidem may be an acceptable
choice, but take into consideration the moderately high
incidence of disordered sleep behaviors while on zolpidem
and similar drugs. Given its short half-life and high seda-
tion effect, low-dose quetiapine may also be an acceptable
short-term solution, especially in patients experiencing
sundowning.
OT H E R AG E N T S T H AT H AV E S HOW N S OM E
PROM I S E I N DE L I R I U M M A N AG E M E N T
Some have theorized that an impaired serotonin metabolism
may play a role in the development of delirium (Maldonado,
2008a,b). The original open-label study found that the anti-
emetic agent ondansetron, a selective serotonin 5-HT3-type
receptor antagonist, may be effective in the treatment of
“hyperactive delirium” (P = 0.001) (Bayindir, Guden Akpinar,
Sanisoglu, & Sagbas, 2001). More recently, there was a second
study (DBRCT; n = 80) comparing ondansetron (8 mg IV)
to haloperidol (5 mg IV) in the treatment of post-cardiotomy
“delirium.” The authors reported no significant baseline dif-
ference between the groups and similar clinical improve-
ment, from delirious baseline, in both groups (p <0.01).
Unfortunately, this study did not use a standardized and
validated delirium severity measure. The nonvalidated “scale”
used by these researchers (different in each case) appeared
to have been more a nonstandardized measure of agitation,
rather than delirium. It is also not clear how the groups were
randomly assessed, and it is difficult to make sense pharma-
cologically of the time for assessment, given very different
pharmacokinetics of both agents (e.g., after IV administra-
tion, haloperidol’s half-life is 10–20 hrs., vs. ondansetron’s 5
hrs.; onset of action is also expected to be quite varied). Given
the nature of these studies, and the fact it did not rely on the
use of a validated diagnostic method, further, better designed
studies are needed.
41. D E L I R I U M
Finally, an open-label, prospective NRCT of patients
undergoing cardiac surgery compared placebo versus statin
administration for prevention of delirium in elderly patients
(Katznelson et  al., 2009). The authors remarked that statin
administration has been shown to decrease morbidity and
mortality after cardiac and major noncardiac surgery, and
that animal and human studies have demonstrated the ben-
eficial effect of statins in central neural system injury. Thus
they theorized that the antithrombotic, antiinflammatory,
and immunomodulatory properties of statins may be respon-
sible for these protective effects, and that the same qualities
may help protect against delirium. In fact, their study dem-
onstrated that the administration of statins had a signifi-
cant protective effect, reducing the odds of delirium by 46%
among patients 60  years old and older. Of interest, these
delirium-sparing effects did not manifest in patients younger
than 60 years old.
DE L I R I U M M A N AG E M E N T: S U M M A RY
In summary, in order to prevent delirium, we recommend
incorporating nonpharmacological and pharmacological
approaches. Key among the possible interventions is the mini-
mization of prolonged and uninterrupted sedation, by sched-
uling daily sedation holidays and titration of sedation to a
targeted goal, assisted by the use of sedation scales. The choice
of sedative and analgesic agents may also make a significant
difference. It is important to be mindful of the additive effect
that certain pharmacological agents may have, including
sedative, deliriogenic, and anticholinergic potential. A good
prevention strategy is the routine inspection of the patient’s
medication list, wise choice of agents (avoiding anticholiner-
gic and benzodiazepine agents, if possible) and eliminating
all unnecessary agents. Remember the rule:  the greater the
number of pharmacological agents the patient is exposed to,
the greater the risk for delirium. The implementation of delir-
ium surveillance methods (such as any of the widely available
delirium detection tools) enables the early recognition of
delirium, and sometimes the detection of subsyndromal forms
of delirium, allowing for prompt action before a full-blown
syndrome has developed. Early mobilization is paramount.
As a rule, the more vertical and restriction-free a patient is,
the lower the incidence of delirium. It is important to elimi-
nate restrictive devices and to reduce or eliminate sedation
and analgesia as soon as medically appropriate and safe, and
to promote getting patients mobilized and out of bed as soon
as possible. Once they are capable of leaving the bed, patients
should be encouraged to eat all meals sitting at the bedside
chair rather than in bed. The implementation of a multicom-
ponent, nonpharmacological protocol (especially with train-
ing and involvement of family and friends) may make a big
difference for many patients, their family, and hospital staff.
When safe, family and staff should be encouraged to get the
patient out of the unit and to see that the patient receives as
much direct sunlight as possible. This will assist the process of
normalizing the sleep–wake cycle naturally by manipulating
the amount of ambient light. When this is not possible, con-
sider the use of melatonin or a melatonin-agonist, as well as
• 891
any of a whole host of nonpharmacological sleep-promoting
interventions (e.g., avoid scheduling unnecessary blood draws
and vital sign checks in the middle of the night [if the patient’s
medical condition allows it], minimize ambient noise and
unnecessary light exposure, limit the amount of TV-watching
after a certain time of day, and offer ear plugs and eye masks
to enhance the patient’s comfort). Melatonin and melatonin
agonists may offer more benefit and fewer side effects than
conventional sedative agents. The selective use of prophylac-
tic antipsychotic agents should be considered, particularly in
patients at high risk or those with a prior history of develop-
ing delirium in the medical environment. When delirium has
already developed, the judicious use of dopamine antagonist
agents seems to have the most evidence for a prompt recov-
ery. When it comes to the use of acetylcholinesterase inhibi-
tors, the data available seem to suggest that they work only in
patients who have been on them long-term (i.e., for dementia).
Thus their use may be considered in patients with moderate to
severe cognitive impairment who suffer from multiple medi-
cal and chronic medical issues and who are expected to repeat-
edly return to the medical environment. Yet, data suggest
that physostigmine is an underutilized therapeutic option we
should consider in obvious cases of anticholinergic delirium.
S U M M A RY
Delirium is a neurobehavioral syndrome caused by the tran-
sient disruption of normal neuronal activity secondary to
systemic disturbances, and is the most common psychiatric
syndrome found in the general hospital setting. In addition
to causing distress to patients, families, and medical caregiv-
ers, the development of delirium has been associated with
increased morbidity and mortality, increased cost of care,
increased hospital-acquired complications, poor functional
and cognitive recovery, decreased quality of life, prolonged
hospital stays, and increased placement in specialized inter-
mediate and long-term care facilities.
Clinical experience and research data suggest that once
delirium has occurred, it is possible the patients may not
return to their pre-delirium cognitive functional level. In fact,
data suggests delirium the occurrence of delirium is associated
with an increased risk of dementia, and acceleration in the rate
of cognitive decline in those demented patients who develop
delirium. Therefore, given increasing evidence that delirium
is not always reversible, and given the potentially numerous
negative sequelae associated with its development, physicians
must do everything possible to prevent its occurrence.
The data also suggest that there may be a “dose-effect” for
delirium, meaning that the longer a patient is delirious, the
worse is the expected outcome. This is particularly true of
hypoactive delirium, which unfortunately is the most com-
monly occurring type, particularly in the elderly. Therefore,
clinicians are encouraged to implement as many prevention
strategies as are available to them in order to improve patient
outcome and prevent or minimize the occurrence of delirium,
thus minimizing its morbidity.
892 • P S YC H I AT R I C C A R E O F T H E M E D I C A L PAT I E N T
Because delirium is common, it is important that physi-
cians be mindful and implement surveillance and monitoring
strategies to allow for early detection and the implementation
of techniques that may shorten its duration. When delirium
occurs, it is important to immediately correct the underlying
contributing causes and use treatment strategies directed at
preventing harm to the patient and others and to shorten the
duration of the delirium episode, thus improving the odds of
recovering baseline functional status.
The psychosomatic medicine specialist has much to con-
tribute to the diagnosis, treatment, and study of delirium. In
the author’s experience, better results at implementing change,
fostering psychiatric consultation, and greater involvement of
our services come from getting involved in multidisciplinary
quality-improvement projects that involve physicians, nurses,
and ancillary staff in the process of recognition and treat-
ment of delirious patients. Active participation in educational
programs (e.g., continuing education, grand rounds, clinical
case conferences) and development of a multidisciplinary
task force charged with assessing the problem and develop-
ing an action plan are key to being recognized as an expert
in the field, the one to go to for advice. At our institution, we
developed a “CNS Pharmacotherapy Task-Force” that devel-
oped protocols for the ICU in sedation, pain management,
neurological assessment, and delirium monitoring, preven-
tion, and treatment. Its success has led to this model being
replicated throughout the institution, with a greater recogni-
tion of delirium as a problem and of psychosomatic medicine
specialists as the experts in the field. The author’s team has
implemented regular didactic sessions for nursing, regular
rounding sessions in the MICU and SICU, and a monthly
“psychiatric mortality and morbidity” review in the internal
medicine program that has been invaluable in helping the
medical staff change their perception and practice toward
delirium, leading to greater joint participation between psy-
chiatry and medico-surgical services.
C L I N IC A L   PE A R L S
• Four important non-modifiable factors for delirium
are older age, baseline cognitive impairment, severity
of underlying medical illness, and preexisting mental
disorders.
• Among patients with preexisting mental disorders,
patients with bipolar disorder had the highest incidence
of delirium.
• Opioids, corticosteroids, and benzodiazepines are major
contributors to delirium see Figure 41.7 (Gaudreau,
Gagnon, et al., 2005b).
• A study of adult patients admitted to inpatient medicine
wards showed that those who developed delirium had
significantly elevated levels (i.e., above the detection
limit) of IL-6 (53% versus 31%) and IL-8 (45% versus
22%), compared with patients who did not develop delir-
ium, even after adjusting for infection, age, and cognitive