The phenotype of delirium based on a close reading of diagnostic criteria

Abstract

Objective

Although delirium is well known to acute care clinicians, the features required for its diagnosis and how to understand and operationalize them remain sticking points in the field. To clarify the delirium phenotype, we present a close reading of past and current sets of delirium diagnostic criteria.

Methods

We first differentiate the delirium syndrome (i.e., features evaluated at bedside) from additional criteria required for diagnosis. Next, we align related features across diagnostic systems and examine them in context to determine intent. Where criteria are ambiguous, we review common delirium instruments to illustrate how they have been interpreted.

Results

An acute disturbance in attention is universally attested across diagnostic systems. A second core feature denotes confusion and has been included across systems as disturbance in awareness, impaired consciousness, and thought disorganization. This feature may be better understood as a disturbance in thought clarity and operationalized in terms of neuropsychological domains thereby clearly linking it to global neurocognitive disturbance. Altered level of activity describes a third core feature, including motor and sleep/wake cycle disturbances. Excluding stupor (wherein mental content cannot be assessed due to reduced arousal) from delirium, as in DSM-5-TR, is appropriate for a psychiatric diagnosis, but the brain injury exclusion in ICD-11 is unjustified.

Conclusions

The delirium phenotype involves a disturbance in attention, qualitative thought clarity, and quantitative activity level, including in relation to expected sleep/wake cycles. Future diagnostic systems should include a severity threshold and specify that delirium diagnosis refers to a 24-hour period.

Introduction

What does delirium look like? After decades of intent delirium research, the answer to this basic question is not nearly as clear as one might expect. Clinical and scholarly interest in this condition has expanded far beyond psychiatry, and delirium is routinely diagnosed by a wide range of clinicians. However, this welcome infusion of perspectives on delirium makes having a clear clinical definition even more important. Further, the delirium phenotype is often complex and difficult to differentiate from conditions that present similarly,¹ which calls for the clinical skill of psychiatrists and neuropsychologists who are expert in performing mental status examination in acute settings.²

Many factors currently conspire against a shared understanding of the delirium phenotype. These include specialty- and discipline-specific silos and training traditions,³ the use of varied terminology,⁴ and a proliferation of delirium instruments for clinical and research use that often differ substantively from one another.⁵ In response, one might appeal to diagnostic criteria as the reference standard, yet even here, at the level where delirium diagnosis should be unambiguous, one finds a deeper layer of potential confusion regarding how to interpret or operationalize individual criteria.

Since the third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III) offered the first widely used set of diagnostic criteria for delirium in 1980, each subsequent definition in the DSM and International Classification of Diseases (ICD) has differed from its predecessors with only one exception (i.e., DSM-IV-TR). Some of these changes have been minor. However, several have introduced stark discontinuities regarding the definition of delirium, and the rationales for most of these changes are not public record. Such changes have introduced uncertainty into the research field as well, both in terms of how to interpret new language and how to operationalize features.

Whereas changes in diagnostic criteria are generally expected to lead to some degree of diagnostic discordance,⁶ clinical efforts to keep pace with successive editions of the DSM have inadvertently complicated matters as each edition has given rise to new delirium instruments⁷ and inconsistent approaches to interpreting the wording of specific criteria.^6,8 These instruments, then, have been variously adopted across clinical settings, a process requiring substantial implementation efforts. As a result, the current clinical landscape comprises an array of different historical and current criteria distributed and preserved in a patchwork pattern across electronic medical records platforms, hospital protocols, and overlain palimpsests in the personal experiences of individual clinicians.

The current state of confusion regarding delirium diagnosis calls for a systematic survey of delirium diagnostic criteria. We offer a close reading of individual diagnostic criteria to bring clarity to the delirium phenotype. This review examines published diagnostic criteria of delirium to identify throughlines that connect specific delirium features over time, while also exploring important distinctions and omissions.

Materials & Methods

We begin by tabulating published delirium diagnostic criteria to align similar items across sets of criteria. The seven unique DSM or ICD sets of diagnostic criteria for delirium are included,^9–16 as well as the Trzepacz, Meagher, and Franco (TMF) research diagnostic criteria¹⁷ (Table 1). To our knowledge, only the TMF criteria have published empirical support. The criteria in DSM-IV and DSM-IV-TR are combined as they are identical; however, the criteria in DSM-5 and DSM-5-TR are considered separately as the changes between them, though linguistically minor, may have an outsized effect on caseness.

Table 1:

Diagnostic Criteria Across DSM and ICD Editions, Including TMF Research Diagnostic Criteria

		DSM-III(7)	DSM-III-R(8)	ICD-10(9)p58	DSM-IV(10) & IV-TR(11)	DSM-5(12)	DSM-5-TR(13)	ICD-11(14)	TMF Criteria(15)
The syndrome of delirium
Onset over short period		X	X		X	X	X	X	X
Fluctuating severity		N	N		N	N	N	N	N
Duration specifier						Specifiers: acute or persistent	Specifiers: acute or persistent
Boundary with normality (threshold)								X
Disturbance in attention		R shift, focus, sustain	R maintain and shift	R direct, focus, sustain, shift	R focus, sustain, shift	R direct, focus, sustain, shift	R direct, focus, sustain, shift	R “a disturbance”	R “impaired ability”
Disturbance in awareness (variously described as consciousness)a		Ra	Omitteda (see disorganized thinking)	Ra	Ra	Ra	Ra	Ra	Omitted, but includes higher-level deficits evidenced by both (1) & (2):
Disorientation		R if testable	P (2 of 6)	R global disturbance of cognition (a disturbance in any of these domains appears to qualify)	Re any change in cognition or perceptual disturbance	Re any disturbance in cognition in addition to awareness and attention	Re any disturbance in cognition in addition to awareness and attention	R “a disturbance”	(1) deficit in at least one domain (in addition to attention)
Memory impairment		R if testable	P (2 of 6)
Disorganized thinking		P (2 of 4)	R “disorganized thinking…indicated by…speech”						(2) thinking, comprehension, or communication
Perceptual disturbance		P (2 of 4)	P (2 of 6)						N
Disturbance of sleep/wake		Pb1 (2 of 4)	Pb1 (2 of 6)	Rb2					R “circadian abnormalities”
Increased psychomotor activity		P (2 of 4) ↑ or ↓	P (2 of 6)	R ↑, ↓, or shifts		Specifier: hyperactive	Specifier: hyperactive
Decreased psychomotor activity			P (2 of 6)			Specifier: hypoactive	Specifier: hypoactive
Mixed level of activity						Specifier: mixed level of activityf	Specifier: mixed level of activityf
Emotional disturbances				Rd					N
Behavioral disturbances
Criteria for diagnosing the disorder of delirium
Exclusions	Dementia				preexisting, established, or evolving “dementia”	preexisting, established, or evolving “neurocognitive disorder”	preexisting, established, or evolving “neurocognitive disorder”	pre-existing or evolving “neurocognitive disorder”	neurocognitive disorder (dementia or mild cognitive impairment)
	Another mental health disorder							X
	Stupor/coma					severely reduced level of arousalg	severely reduced level of arousalg		coma, stupor, sedation
	Substance intoxication or withdrawal							typical syndrome (intoxication or withdrawal)
	Acquired or traumatic brain injury							preceding neurological injury or event
Physiological attribution		X evidence (history, exam, or labs)	X evidence (history, exam, or labs) OR presumed cause		X evidence (history, exam, or labs)	X evidence (history, exam, or labs)	X evidence (history, exam, or labs)	X qualifying evidence unspecified	X “temporally related to one or more identifiable physiological or pharmacological potential etiologies”
Medical condition		organic factor	organic factor	“not induced by alcohol and other psychoactive substances”	general medical condition	another medical condition	another medical condition	medical condition
Withdrawal	Alcohol	Xc	Xc	X (specifier for withdrawal)	X	X	X	X individual agents not specified
	Opioid					X	X
	Sedative-hypnotic	Xc	Xc		X	X	X
	Other/unknown	X	X		X	X	X
Intoxication	Alcohol			X (specifier for intoxication)	X	X	X	X individual agents not specified
	Amphetamine	X	X		X	X	X
	Cannabis				X	X	X
	Cocaine				X	X	X
	Hallucinogen				X	X	X
	Inhalant				X	X	X
	Opioid				X	X	X
	Phencyclidine	X	X		X	X	X
	Sedative-hypnotic				X	X	X
	Other/unknown	X	X		X	X	X
Medication-induced					as “Substance Intoxication”	as “side effect”	as “side effect”	X
Multiple etiologies				included as Other delirium (F05.8)	X	X	X	X
Unspecified or not otherwise specified			(cause can be presumed, see Physiological attribution above)	delirium, unspecified (F05.8)	insufficient evidence of cause or cause not listed (e.g., sensory deprivation)	symptoms of delirium that cause distress or impairment but do not meet full criteria; clinician chooses not to specify the reason criteria are not met	symptoms of delirium that cause distress or impairment but do not meet full criteria; clinician chooses not to specify the reason criteria are not met	cause is unknown or unspecified
Other specified				other delirium (F05.8)		symptoms of delirium that cause distress or impairment but do not meet full criteria; clinician specifies (e.g. attenuated delirium syndrome)	symptoms of delirium that cause distress or impairment but do not meet full criteria; clinician specifies (e.g. subsyndromal delirium)	specified cause is known to be capable of causing delirium
Delirium superimposed on dementia			coding note	delirium, superimposed on dementia (F05.1)	coding note	mentioned only in text	mentioned only in text	coding note

Abbreviations: P, polythetic; R, required; X, included; N, described but non-diagnostic; blank cell, omitted from diagnostic system

^aSee Table 3 for details

^b1Insomnia or daytime drowsiness;

^b2Insomnia, total sleep loss, sleep-wake cycle reversal, daytime drowsiness, nocturnal worsening of symptoms, disturbing dreams, or nightmares

^cRequires autonomic hyperactivity

^dExamples include depression, anxiety, fear, irritability, euphoria, apathy, or wondering perplexity

^eAt least one additional item required, though a complete list of qualifying cognitive domains is not provided

^fNormal psychomotor activity or psychomotor activity rapidly fluctuates

^gThe text of DSM-5 specifies that minimal responses to verbal stimulation should be classified as severe inattention whereas the text of DSM-5-TR specifies that minimal responses to verbal or physical stimulation should be classified as a disorder of arousal such as coma or stupor, not as delirium

Each feature is examined for its unifying theme, apparent intent based on wording and context, and potential operationalization. Where diagnostic criteria are ambiguous, we review common instruments to illustrate how these items have been interpreted; however, a systematic survey of delirium instruments is beyond the scope of this paper.¹⁸ Diagnostic criteria are considered in two sections: criteria that characterize the syndrome of delirium (i.e., features encountered at bedside) vs the broader delirium disorder (i.e., additional criteria required for diagnosis).¹⁹ Based on a close reading of individual criteria, we offer insights on how best to understand individual items and to highlight areas of needed development.

Results

Section 1: The syndrome of delirium

Natural history

The syndrome of delirium represents a marked change in neurocognition that develops over a short period, usually hours to days; therefore, knowledge of a person’s baseline cognitive status is key to accurate diagnosis. The fluctuating nature of delirium—often described as a “waxing and waning” course—is described across systems as common but not required diagnostically. DSM-5 and DSM-5-TR are unique among systems to include course specifiers as either acute “(lasting a few hours or days)” or persistent “(lasting weeks or months)”. Absent in all systems is the period over which a diagnosis is to be made. In clinical and research settings, 24 hours is often presumed, which also allows for describing circadian course.

Threshold (clinical significance)

ICD-11 is unique in describing a “boundary with normality,” emphasizing that delirium should be characterized by “the presence of significant confusion and/or global neurocognitive impairment.” The DSM does not include a clinical significance criterion as it does with other disorders—namely, that the disturbance be required to cause “clinically significant distress or impairment in…functioning.”²⁰

Disturbance in attention

Included in every set of criteria is a disturbance in attention, likely owing to the seminal work of neurologists Chédru, Geschwind, and Mesulam.^21–23 In neuropsychology, attention is understood as a multidimensional construct. Additionally, tests of attention seldom evaluate attention alone; rather, the assessment of attention evaluates attention as integrated with other cognitive networks such as those subserving arousal and executive control. A person with intact attention might perform poorly on tasks commonly thought of as “tests of attention” due to impairments in these other domains.²⁴ For instance, serial 7s and months of the year backward require working memory, whereas the vigilance A task is a relatively pure test of sustained attention (Table 2). Performance on tests of attention, as with most tests, can also be affected by personal factors including emotional state, fatigue, or degree of motivation.

Table 2:

Types of attention

Type	Description	Clinical notes on assessment
Directed attentiona	Voluntary control over attention, as opposed to cued attention	An inability to direct attention by internal control may be evidenced by: an inability to engage in clinical evaluation, haphazard engagement with internal and external stimuli, being entirely stimulus bound, or exhibiting no meaningful attention whatsoever.
Focuseda and selective attention (attentional control, search, or concentration)	Ability to attend to important stimuli or ideas (focused), while suppressing competing stimuli (selective); each requires directed attention	Focused attention also implies some degree of selective attention as internal and external stimuli are constantly vying for attention. Focused attention can be assessed with visual scanning tasks. Selective attention is evaluated in paradigms that present varying types and degrees of competing stimuli, such as Stroop or Eriksen flanker task. At bedside, focused/selective attention may be assessed by asking the patient to direct attention to specific environmental stimuli or to perform (i.e., focus on) cognitive tasks.
Sustained attention (vigilance)	Ability to maintain (directed, focused, or selective) attention over time	Sustained attention requires intact directed, focused, and selective attention.Sustained attention may be evaluated using: Vigilance A task (variations included on CAM-ICU and Montreal Cognitive Assessment) Months of year backward, days of week backward, serial 7’s, serial 3’sb Cancellation test Choice reaction task Continuous performance task
Shifting attention (alternating attention)	Ability to disengage from a current stimulus or idea and redirect attention to another	Shifting attention may be evidenced by verbal or topical perseveration, either of which demonstrates an inability to shift attention to new topics or stimuli. Shifting attention may be assessed with tests of attentional orienting where the examiner provides visual or verbal cues that require the patient to shift attention among them.
Divided attention (tracking)	Ability to respond to more than one task or multiple elements of a task at a time	Divided attention is a higher order form of attention and represents a combination of other types of attention. It may be assessed by: Trail making A & B (sustained and shifting attention; B also assesses executive function) Vigilance task with multiple targets (sustained)
Attention span (attention capacity)	Amount of information that can be grasped at once	Attention span describes an amount of information and implies directed/focused attention. Tests that require the manipulation of information also involve working memory. Tests include: Forward digit span Reverse digit spanb Spatial span

^aDirected and focused attention overlap both conceptually and in literature descriptions.

^bThese tasks assess both sustained attention as well as working memory as they require manipulation of information

The types of attention that satisfy this criterion have varied across diagnostic systems, ranging from “a disturbance” in ICD-11 and reduced ability to “maintain” and “shift” attention in DSM-III-R to “reduced ability to focus, sustain, or shift attention” in DSM-IV-TR and “reduced ability to direct, focus, sustain, and shift attention” in DSM-5. The fact is that these granular distinctions are not only lost on most clinicians without neuropsychological expertise but also so nuanced that common bedside tests are unlikely to differentiate cleanly among them, especially when being assessed in acute care settings where interruptions such as environmental noises and parades of clinicians going in and out of rooms are commonplace. One also presumes that in the DSM-5 and DSM-5-TR criterion for attention (“direct, focus, sustain, and shift attention”), the authors meant “or” rather than “and,” lest the criterion be unintentionally restrictive. Finally, whereas the term inattention is usually used to describe this feature, the term dysattention accurately conveys the intended meaning of varied degrees of disturbance.

Given that attention is often grossly disrupted in delirium, there is limited value in traditional neuropsychological tests of attention because patients with delirium are seldom able to participate in these. The foundational aspect of attention is directed attention (Table 2), which denotes voluntary attentional control. Focused attention is occasionally described synonymously with directed attention but further describes the ability to consolidate attention on a target. Selective attention is closely related to focused attention but denotes the ability to ignore competing stimuli in favor of the selected stimulus. Sustained attention is focused attention over time. Shifting attention is the ability to redirect and refocus attention on new or different stimuli; it requires not only an ability to focus attention but also to disengage attention from a prior stimulus. Divided attention is a higher form of attention, which also requires some degree of directed, focused, or sustained attention. Attention span is a measure of capacity and describes how much information can be grasped at once, and its assessment also requires sufficiently intact directed and focused attention.

Environmental or external stimuli in attention

DSM-III specified that the deficit in attention must be in relation “to environmental stimuli,” and DSM-III-R included a similar phrase. However, no subsequent diagnostic systems have included this qualification. One presumes that a disturbance in attention may be either to environmental (external) or internal stimuli, including mental experiences such as thoughts or sensory experiences, though attention to internal stimuli must be inferred and, as a result, might be less reliable to assess.

Disturbance in thought clarity

Disturbance in awareness, impaired consciousness and thought disorganization

A close reading of criteria over time would suggest that disturbance in awareness (DSM-5, DSM-5-TR, and ICD-11), impaired consciousness (DSM-III, DSM-IV, DSM-IV-TR, and ICD-10), and thought disorganization (DSM-III-R and TMF) intend to describe the same concept related to impaired clarity of thought. Nevertheless, it is striking that this feature, which speaks to the essence of delirium as a confusional state, has met with such varied definitions over time. Although definitions of “confusion” have ranged widely,²⁵ a disturbance in thought clarity is often a central feature. For instance, Victor and Adams described confusion as “an incapacity to think with customary speed, clarity, and coherence”²⁶ and Mesulam and Geschwind as “an inability to maintain a coherent stream of thought and behavior.”²³ Because there is greater linguistic confusion regarding this finding and, as a result, a higher likelihood of misunderstanding, we offer a more detailed analysis in this section.

Disturbance in awareness captures the qualitative aspect of consciousness: it ranges from clear to clouded. Through DSM-IV-TR this criterion was described using the qualifiers “clouding” or “reduced clarity” (Table 3). Unfortunately, using the term “reduced” in combination with “clarity” is nettlesome as this pair of terms combines the quantitative aspect of consciousness (i.e., “reduced” as in “reduced arousal”) with its qualitative aspect (i.e., clarity of the content of consciousness). Given that both hyperactive and hypoactive variants of delirium are acknowledged across all editions of the DSM and ICD, the word “reduced” in relation to “clarity” cannot be referring to arousal. Even worse, ICD-10 describes a spectrum of impaired consciousness ranging from clouding to coma, positing a qualitative term at one end (“clouding”) and a quantitative term on the other (“coma”). One presumes a heteroschedastic relationship between awareness (qualitative) and arousal (quantitative) such that a person with severely reduced or elevated arousal is far more likely to have clouded awareness as well, whereas a person with healthy, waking arousal is far more likely to have clarity of awareness.²⁷

Table 3:

Awareness and Consciousness Across DSM and ICD Editions

	Qualifier	Core domain	Object
DSM-III (1980)	Clouding of	Consciousness
	(Reduced clarity of	Awareness	Of environment)
ICD-10 (1993)	Impairment in	Consciousness
	(Clouding to coma)
DSM-IV & IV-TR (1994 & 2000)	Disturbance of	Consciousness
	(Reduced clarity of	Awareness	Of environment)
DSM-5 (2013)	Disturbance in	Awareness
	(Reduced	Orientation	To environment)
DSM-5-TR (2022)	Reduced	Awareness	Of environment
ICD-11 (2022)	Disturbance in	Awareness

DSM-5’s parenthetical description of the awareness criterion represented a serious departure from all other definitions, preceding and subsequent, by equating disturbance in awareness with disorientation. Attempts at incorporating different interpretations of this wording substantially affected caseness.^6,8 Even a cursory reading of DSM-5 indicates that it could not have been referring to simple orientation because a subsequent criterion lists “disorientation” among the “additional” qualifying features. One presumes DSM-5’s parenthetical use of “orientation” was meant to describe orientation “to situation,” as in being oriented to person, place, time, and situation. Regardless, DSM-5-TR corrected this issue by omitting the misleading term “orientation.” Like DSM-5, DSM-5-TR also excluded reference to “consciousness” in this criterion but instead included “consciousness” in the descriptive text, thereby limiting the chances that the term might be erroneously equated with arousal. ICD-11 includes only a parsimonious awareness criterion, describing it merely as a “disturbance in awareness” without explication.

Although the term awareness appears in the latest DSM and ICD editions, most delirium instruments do not include an explicit “awareness” item, which means that, if an awareness item is being included, it must be under different terminology. An early instance of “awareness” was its parenthetical inclusion on the Memorial Delirium Assessment Scale (MDAS) in the item “reduced level of consciousness (awareness).”²⁸ The fact that the MDAS also included separate items for “decreased or increased psychomotor activity” and for “sleep-wake cycle disturbances (disorder of arousal)” indicates that the awareness item was properly focused on the qualitative aspect of consciousness even though it uses the quantitative term “reduced level” as a synonym in the name of this item. A similar item “Reduced awareness of surroundings” is included on the Delirium Motor Subtype Scale (DMSS) as a feature specific to hypoactive delirium.²⁹ However, the fact that the DMSS includes reduced awareness only in relation to hypoactive delirium seems inconsistent with the DSM because the DSM has required a disturbance in awareness for delirium diagnosis since the 4^th edition yet has also consistently acknowledged both hyperactive and hypoactive subtypes.

The incorporation of this feature as disorganized thinking in DSM-III-R (n.b., the DSM edition from which the Confusion Assessment Method [CAM] was derived³⁰) and in the TMF research diagnostic criteria is far more accessible and intuitive. This interpretation is further supported by the factor analyses that led to the creation of the TMF criteria.³¹ Awareness involves the processing of information clearly and coherently, so one assesses awareness by way of a patient’s narrative responses, their responses to narrative instructions, or—occasionally—as inferred by observing a patient’s engagement with the environment. Informally speaking, the clinician wants to know whether the patient “gets it,” whether they can make sense of “what’s going on” at the time of the assessment.

Disorganized thinking is emblematic of delirium and bespeaks of general confusion. However, disorganized thinking has been operationalized in many ways, reflecting its myriad clinical expressions. The term “disorganized thinking” occurs in DSM-III-R and TMF, as feature 4 on the short-form CAM and CAM for the Intensive Care Unit (CAM-ICU), and item 6 on MDAS. Item 6 on DRS-R98—“thought process abnormalities”—appears to evaluate the same construct.³² Despite the superficial similarity of disorganized thinking across scales, their definitions differ in terms of what qualifies and in terms of which neuropsychological domains are affected (Table 4). Nevertheless, each of the following disturbances denotes confusion or evidence of impaired awareness, be it situational awareness, conceptual awareness, or awareness of interpersonal cues or instructions.

1. Speech amount or clarity: Of the three descriptors of speech listed in Table 4, column 1, “less meaningful and coherent,” “irrelevant” and “incoherent” clearly refer to speech content. In irrelevant speech, the content pertains to topics unrelated or inappropriate to context; in incoherent speech the clarity is lost. Given that these descriptors lead to impaired communication of thought content, they provide proper evidence of thought disorganization. These may be considered types of “clouding of speech clarity.” On the other hand, the term “rambling” is nonspecific. For instance, a person who is enthusiastic or nervous may ramble, so the poor specificity of “rambling” would suggest that it alone, without consideration of speech content, would not necessarily indicate thought disorganization.

2. Thought process: The definitions of qualifying features range from non-specific changes in thought flow (“unpredictable changing of subjects,” “tangentiality,” or “circumstantiality”) to a loss of intelligibility (“unclear or illogical flow of ideas,” “loose associations,” or “impaired comprehension”). For clinical significance, one expects that a diagnostically sufficient disturbance of thought process impairs intelligibility. The most direct means of assessing thought process is based on narrative speech clarity, as above, though it may also be assessed based on responses to instructions.

3. Praxis: The CAM-ICU, which was designed to assess delirium in intubated and therefore non-speaking patients, uses assessments that do not require spoken responses. Giving motor instructions assesses the patient’s comprehension of the verbal content provided and their awareness that a response is being requested. It also requires the ability to perform the task in question.

4. Reasoning: MDAS refers to “faulty reasoning,” and, similarly, the CAM-ICU includes four yes/no questions as tests of basic reasoning. In neuropsychology, reasoning is a unique domain that can be assessed verbally (e.g., word puzzles or math problems) or visually (e.g., completing or arranging pictures). A closely related neuropsychological domain is concept formation, which also can be assessed verbally (e.g., proverbs, similarities, or differences) or visually (e.g., matrices, sorting, or shifting). Responding to tests of reasoning or concept formation requires awareness of the instructions and sufficient clarity of thought to complete the task correctly. It also requires task-specific domains such as language, mathematics, or visuospatial ability.

Table 4:

Definitions of Disorganized Thinking

	Speech amount or clarity	Thought process	Praxis	Reasoning
DSM-III-R	Rambling, irrelevant, or incoherent speech
TMF	Less meaningful and coherent speech	Impaired ability to comprehend
CAM item 4	Rambling, irrelevant, or incoherent speech	Unpredictable changing of subjects; unclear or illogical flow of ideas
CAM-ICU item 4			Praxis	Reasoning
MDAS item 6	Rambling, irrelevant, or incoherent speech	Tangential or circumstantial thought process		Faulty reasoning
DRS-R98 item 6		Tangential or circumstantial thought process; loose associations

Environmental or external stimuli in awareness

The awareness criterion also includes the phrase “of [or to] environment,” which has appeared in every edition of the DSM that includes an awareness criterion but not in ICD-10 or ICD-11. Given that disturbance in awareness signals an underlying disturbance in thought clarity, one presumes that it need not be restricted to the surrounding environment. A person may similarly have a disturbance in awareness of their broader situation (e.g., social, physical, medical) or to internal stimuli (as in the historical “clouding of sensorium”) although we are not aware of prior attempts at operationalizing a disturbance in awareness as such, as it is likely difficult to measure these reliably.

Neurocognitive disturbance

Every system requires some degree of neurocognitive disturbance, but they differ in terms of which disturbances are required. DSM-III and DSM-III-R identified individual cognitive domains, each with its unique set of required or polythetic options. Taking a different approach, ICD-10 and DSM-IV through 5-TR combined all cognitive disturbances as a single set and have allowed any disturbance to satisfy this criterion—that is, in addition to a disturbance in attention and awareness.

Disorientation

Evaluating orientation is usually straightforward. Disorientation to time typically occurs first in delirium, followed by place and person.^{25 [p93]} Certain elements of disorientation may be difficult to assess in some contexts, as when this information is posted in clear view of the patient who can simply read it or encode the correct information and then perseverate on the correct answer throughout the assessment. Although delirium commonly presents with some degree of disorientation, both DSM-III and ICD-11 require disorientation for diagnosis.³³ Further, disorientation is not required on the most commonly used bedside delirium assessments (e.g., orientation is included in the long-form but not the short-form CAM; similarly, it is part of the 4-AT but not required for a positive screen³⁴).

Learning and Memory

Memory is a multidimensional construct ranging from immediate (seconds) to long-term (decades) and involves several steps (e.g., registration, retention, and retrieval).³⁵ Memory assessments in acute settings are often limited due to degree of patient engagement and a variety of care- and context-related factors. Most memory assessments in delirium involve verbal content, pertaining to either semantic or episodic memory; however, impaired visuospatial or autobiographical memory would also qualify. Implicit memory is rarely considered in delirium, though it has not been studied systematically. Learning new information requires information first be registered, which requires adequate attention, so impaired learning in delirium is often the result of impaired attention.³⁶ Similarly, because delirium affects sustained attention, impairments in retrieval may similarly be attributable to a person’s inability to conduct a sustained search of previously registered information.

Language impairment

Although a patient with delirium is seldom able to express themselves cogently, this is rarely the result of a primary impairment in language. For instance, word finding difficulty or impaired reading or writing may be the result of impaired attention, concept formation, episodic memory, or executive function. It is uncommon for patients to exhibit formal language deficits such as dysfluent speech, impaired comprehension, or agrammatism. What one typically encounters is a disturbance in thought clarity, as above.

Visuospatial/constructional performance

These two domains are often considered in tandem because construction—the ability to draw objects, build, or assemble—requires intact visuospatial perception and the ability to orient or organize oneself in space. The relationship between visuospatial ability and constructional performance is akin to the relationship between thoughts and speech as the prior is often inferred from the latter. Unfortunately, although visual constructional ability is generally impaired in delirium,³⁷ it may have limited ability to discriminate delirium from dementia.³⁸ Whereas performance on certain visuoperceptual tasks (e.g., dot counting, shape detection, or identifying incomplete letters) may be more impaired in delirium than in dementia,³⁹ it remains unclear to what extent such performance impairment may be explained by deficits in attention or other basic elements of consciousness. The clock-drawing test may be well-suited for many patients,⁴⁰ though the absence of standards for assessing patients physically incapable of drawing is an unmet clinical need.

Perceptual disturbance

Perceptual disturbances may occur across sensory modalities and, as on the DRS-R98⁴⁰ or CAM,³⁰ include misperceptions, illusions, or hallucinations. In addition, a growing body of literature on the lived experience of delirium offers a better appreciation of its unreality.⁴¹ Fact and fiction can be confused, reality transmogrified, the flow of time upended, and relationships with others, the environment, and even oneself disconnected. Such perceptual disturbances may be best characterized phenomenologically as forms of dissociation. Additionally, whether such experiences represent the intrusion of dream content into wakefulness, as the boundary between the states of sleep and wake dissolve, remains an unanswered question.⁴² Notably, the TMF excludes perceptual disturbances as a qualifying cognitive feature of delirium.

Delusions (and other erroneous beliefs)

Although delusions occur in a minority of patients with delirium,³⁷ they have long been described as a potential delirium feature. The DSM-5-TR definition of delusion applies well to what is often encountered in delirium: fixed beliefs that are not open to change in light conflicting evidence. As with all beliefs, delusion severity can range across several continua (e.g., degree of conviction, degree of encapsulation, bizarreness, organization). The anchor points on the DRS-R98 offer a clinically informed gradient including mild suspicion, overvalued ideas, and delusions.³² Whereas delusions may often appear obvious to the clinician, there are several related syndromes that should be differentiated: confabulation, dissociation, and anosognosia (lack of awareness of impairment). In fact, a patient may deny even the most obvious of physical or functional impairments, including blindness as in Anton syndrome, or act as though these impairments are either not present or, if present, irrelevant. As with perceptual disturbances, delusions are also excluded from the cognitive disturbance item per TMF and are associated with greater delirium severity overall.⁴³

Proposal

Clinically significant disturbances in thought clarity will lead to impairment in neurocognition. We propose that future diagnostic systems consider incorporating awareness as disturbance in thought clarity and formulating its assessment in terms of specific neuropsychological domains—for instance, as “a disturbance in the clarity of thinking as evidenced by any of the following: an inability to communicate coherently, an inability to understand verbal questions, disorientation to situation, an inability to register or recall information, an inability to solve simple questions of logic, or an inability to follow simple motor commands [provided any such findings are not better explained by aphasia].” These items correspond to expressive and receptive language, orientation (situational awareness), learning/memory, executive function, and praxis, respectively. Explicitly linking thought clarity—i.e., “awareness” in DSM-5-TR and ICD-11—with clinical interaction and neurocognitive performance is a clinically intuitive way of understanding “confusion” and could also be mapped onto the TMF and CAM-family of assessments.

Disturbance in activity level

Sleep/wake cycle disturbance

The canonical states of sleep and wake are inherently disrupted in delirium: an apparently waking patient exhibits diffuse slowing on electroencephalography as occurs in non-REM sleep, and an apparently sleeping patient often lack common sleep-phase–defining transients,⁴⁴ indicating the absence of physiological sleep phases. A sleep/wake disturbance was included as a polythetic option in DSM-III and DSM-III-R and required for diagnosis in ICD-10, yet the current editions of DSM and ICD do not include this item. Delirium and sleep/wake disturbances may involve interrelated neurophysiological processes including the basal forebrain, diencephalon, and the reticular activating system.^45,46 Whereas clinicians have recognized the association of delirium and sleep/wake disturbances for millennia,²⁵ it remains unclear to what extent such disturbances represent the effects of change in arousal, changes in brain networks that subserve sleep phases, or a disturbance in endogenous circadian rhythms. TMF criteria require either a disturbance in sleep/wake or a change in motor activity for diagnosis, as these two features load onto the same latent factor.³¹

Change in psychomotor/motor activity

DSM-III and DSM-III-R included changes in activity level as polythetic options. ICD-10 required behavioral changes for diagnosis; and DSM-5 and DSM-5-TR include behavioral subtypes as specifiers: hyperactive, hypoactive, or mixed level of activity (including both normal and rapidly fluctuating levels of activity). The inclusion of a “normal” level of activity in “mixed level of activity” is especially curious as a purely “normoactive delirium” has not universally been included as a potential delirium variant, though periods of normoactivity may be expected during transitions between motoric poles.

Motor activity refers to global activity level, with some understanding this as a psychomotor phenomenon^11,15,47,48 and others as pure motor^17,29,40 changes. The motoric subtypes of delirium have been examined using accelerometry-based analysis^49–51 leading to validated motion-based criteria,⁵² which has paralleled efforts to develop valid clinical assessments of delirium motor subtypes, including the Delirium Motor Subtyping Scale⁵³ and an abbreviated version.⁵⁴ Many authors consider changes in activity as linked with changes in arousal^36,55 or alertness,^{25 [p44]} which is important because the increased mortality risk associated with delirium^56,57 is driven in part by reduced arousal,⁵⁸ underscoring the importance of including activity changes as a core feature of delirium rather than only as a specifier.

Emotional disturbances

Delirium all but invariably includes some degree of emotional disturbance, yet these features have long escaped systematic evaluation.¹⁹ ICD-11 includes them among its Additional Clinical Features, and the text of DSM-5-TR makes a passing comment about emotional disturbances under the section Associated Features. Further, very few delirium assessments include an item for emotional disturbance.¹⁸ For instance, DRS-R98 includes “lability of affect,” ranging from mild alterations in affect to severe, disinhibited, or rapidly changing affect; however, the emotional disturbances that can occur in delirium are much broader and include, for example, anxiety, fear, depression, apathy, irritability, anger, or euphoria. Such experiences are often among the most distressing elements of delirium based on patient self-reports.^41,59

Recent research has begun to characterize the experiential burden of delirium, including the delirium experience itself, its associated emotional disturbances, and situational distress,⁶⁰ leading to the creation of the Delirium Burden scale for patients and caregivers.⁶¹ Depression, in particular, has been studied as a delirium risk factor⁶² and outcome,⁶³ but its occurrence during delirium is poorly understood. Additionally, the links among the “three D’s” of depression, dementia, and delirium deserve further attention,⁶⁴ as each additional syndrome is associated with incrementally worse functional status.^65,66 Future studies would do well to characterize the clusters of emotional disturbances in delirium analogous to the neuropsychiatric disturbances of dementia.⁶⁷ It also remains unclear the extent to which emotional disturbances are due to specific underlying delirium causes or psychological responses to the delirium experience.

Section 2: Criteria for diagnosing the disorder of delirium

Differential diagnosis

Delirium is a universal rule-out diagnosis in psychiatry that precludes diagnosis of primary psychiatric disorders. This exclusion prioritizes medical workup and clinical management of its cause(s). It is also necessary because delirium can mimic most primary psychiatric disorders. Nevertheless, the fact that delirium should be differentiated from several conditions before being diagnosed is a critical reason why simply identifying the syndrome of delirium (e.g., by way of brief clinical instruments) should not be conflated for diagnosis.¹⁹

Pre-existing, established, or evolving neurocognitive disorder (dementia)

Differentiating delirium from dementia is essential but can be challenging. Delirium should represent a change from baseline, as distinct from pre-existing or established neurocognitive disorders. It is unclear what the term “evolving” means in relation to neurocognitive disorders or dementia in DSM-IV through DSM-5-TR and ICD-11, though one imagines that this would include the hallmark cognitive fluctuations in dementia with Lewy bodies⁶⁸ and the global cognitive changes encountered in rapidly progressive dementias such as prion disease.⁶⁹

Stupor/coma

Whereas excluding coma from delirium diagnosis is unobjectionable, where to draw the line between states of severely reduced arousal and delirium remains controversial. Although unaddressed in the ICD, the two most recent editions of DSM have offered contrasting instructions. The text of DSM-5 explained that “minimal responses to verbal stimulation should be classified as severe inattention, hence delirium,” whereas the text of DSM-5-TR specifies that “minimal responses to verbal or physical stimulation should be classified as a disorder of arousal such as coma or stupor, not as delirium.”

We favor a more restrictive definition of delirium because mental content is essential to this syndrome, lest delirium be reduced to a scale of arousal. It is also safer to treat patients with intermediate states between delirium and coma—states occasionally described as “stupor”—as “near-coma” rather than as delirium because causes of coma are typically more severe and of greater clinical urgency,²⁷ with greater risk of mortality.⁷⁰ Furthermore, including stupor as delirium⁷¹ would admit a form of delirium diagnosis sans delirium, a confusion-less delirium. Ultimately, whereas the psychiatric tradition has been to exclude states of severely reduced arousal from delirium, anyone using an inclusive approach should specify this and distinguish such patients where the content of consciousness cannot be assessed reliably (e.g., “delirium, with stupor,” or similarly) for sub-analyses.

Acquired or traumatic brain injury

Excluding delirium after a preceding neurological event, which is unique to ICD-11, is unjustified as we see no reason why a traumatic or other acute brain injury should be separated from every other pathophysiological insult that leads to the syndrome of delirium,²⁰ especially given that post-traumatic delirium has the same syndromal structure as elsewhere.⁷² The convergent syndrome of delirium is a testament to the unifying value of this diagnosis, a value undermined by such arbitrary exclusions.³

Substance intoxication or withdrawal

ICD-11 specifies that typical states of substance intoxication or withdrawal should not be diagnosed as delirium; however, a diagnosis of delirium is appropriate where the presentation meets diagnostic criteria. DSM specifies that cognitive symptoms should be “in excess”¹³ of typical intoxication or withdrawal syndromes or “predominate in the clinical picture”^14,15 for delirium diagnosis.

Another mental health disorder

This differential, unique to ICD-11, draws attention to primary psychiatric conditions that can mimic the syndrome of delirium and as a result can be misdiagnosed. Psychiatric diagnoses that should be considered on the differential for delirium include depression, catatonia, disorganized psychosis, delirious mania, and Ganser syndrome along with other dissociative disorders.¹

Physiological attribution and potential causes

The lynchpin of delirium diagnosis, as opposed to simply identifying its syndromal features, is attributing the clinical presentation to the direct physiological effects of a medical condition, substance, or a combination of causes, which include the following: (a) Medical condition. It is beyond the scope of this review to list all potential medical causes of delirium, but readers are referred to chapters and systematic reviews on this topic.⁶² (b) Substance withdrawal. The list of substances from which a person may develop delirium in withdrawal is generally limited to sedatives, chiefly those with gamma-aminobutyric acid activity.^15,16 (c) Substance intoxication. The range of substances that may cause delirium in intoxication includes substances of abuse as well as a host of toxins, the listing of which extends far beyond those listed in the most recent editions of the DSM (Table 1). (d) Medication-induced. Prescribed and over-the-counter medications can also cause delirium. For instance, this may be due to anticholinergic effects, the cognitive effects of supratherapeutic levels (e.g., in overdose, intentional or otherwise), or to medication-specific effects (e.g., delirium due to cephalosporins).^15,16 (d) Multiple etiologies. Although a primary cause of delirium is often identified, multiple causes are identified in most cases.²⁹ Therefore, it is important to avoid anchoring bias after finding one potential cause. (e) Unspecified. This diagnosis is important for instances where the syndrome of delirium is present, and the clinician attributes this to an as-yet unidentified pathophysiological precipitant, as when a medical workup is ongoing or where the clinician attributes the syndrome of delirium to a cause that was never formally identified.^15,16 (f) Other specified. This diagnosis describes scenarios where a person presents with an incomplete syndrome of delirium.^15,16

Conclusions

A robust and expanding delirium literature speaks to the growing interest in the field and to the degree of sophistication with which modern investigators are approaching this condition described since antiquity. Several tentative conclusions may be drawn from this analysis (Table 5). The three key disturbances of the delirium syndrome involve attention, the qualitative element of thought clarity, and the quantitative element of activity level. Two criteria that should be included in future systems are: specified 24-hour duration for diagnosis and a severity threshold. Change in activity level or circadian pattern should be re-evaluated for inclusion as a core feature in future diagnostic systems. Delirium in the context of brain injury should be included as a form of delirium, and the field needs to come to consensus on how to designate patients with stupor. Additional research is needed on the broad range of neuropsychiatric disturbances that occur in delirium. Finally, future diagnostic systems should consider the heuristic value in differentiating criteria that describe the delirium syndrome from the additional criteria required for diagnosis.

Table 5:

Areas for further development of the delirium construct

Diagnosis	Structure: there is heuristic value for differentiating criteria that describe the delirium syndrome from the additional criteria required for diagnosing the broader disorder of delirium
Natural history	Period of diagnosis: should be specified as 24 hours
Threshold	Clinical significance: a threshold for clinical significance should be specified
Attention	Domains: descriptions of the types of attention that qualify are suggested
Thought clarity (awareness)	Operationalization: propose defining disturbance in awareness as a disturbance in clarity of thinking as evidenced by impairment in specific neuropsychological domains
Activity	Reclassification: deserves to be considered a core feature of delirium Operationalization: as either change in psychomotor/motor activity or as disturbance in sleep/wake cycle
Neuropsychiatric disturbances	Characterization and impact: the role of neuropsychiatric disturbances such as depression, irritability, and distress in delirium deserve investigation
Differential diagnosis	Dementia: further research on differentiating from cognitive fluctuations of dementia with Lewy bodies and rapidly progressive dementias Stupor/coma: stupor should be designated separately from states where the content of consciousness is reliably evaluable, even if one considers these states a form of severe delirium Traumatic or acquired brain injury: encourage integration of posttraumatic confusional state with delirium as posttraumatic delirium Primary psychiatric disorder: should be added as an exclusion in future diagnostic systems to highlight psychiatric conditions that can mimic delirium

Key points.

1. Disturbance in attention is a hallmark feature of delirium and may manifest in any number of attentional domains (e.g., directed, focused/selective, shifting, sustained, or attention span).

2. The second hallmark feature of delirium, disturbance in thought clarity, has met with considerable confusion and been included variously across diagnostic systems as a disturbance in awareness, impaired consciousness, or thought disorganization.

3. Altered level of activity, either cross-sectionally or in relation to sleep/wake patterns, represents a third core feature of delirium.

4. Excluding stupor from delirium is justified for a psychiatric diagnosis, but acute confusion after brain injury that otherwise meets delirium criteria should be regarded as delirium.