The definition of delirium previously had been considered in simplified terms as an acute confusional state. However, since the 1980s, it has been refined, whereby earlier versions of the Diagnostic and Statistical Manual of Mental Disorders (DSM) were more weighted upon the causes of generalized brain dysfunction than stratification of clinical constructs that define the disorder [12]. The construct and definition of delirium continues to evolve with the recent revision of the DSM [13]; the DSM-5 avoids the term “consciousness” and more restrictively defines delirium in terms of cognitive impairment and inattention with less emphasis on the level of arousal. The DSM-5 criteria (Table 7.4) also take into account established or preexisting neurocognitive disorders. In addition, DSM-5 criteria consider the context of acute cognitive impairment and fluctuation in relation to an individual’s baseline function. However, differences between DSM-5 and DSM-IV criteria have raised concerns of the sensitivity of detecting delirium based on the revised criteria [14]. Hence, recommendation for broader inclusion to include individuals with cognitive impairment secondary to impaired arousal (e.g., drowsiness, obtundation, stupor) as well as consistent interpretation of criteria is thought to be important factors especially with regard to patient safety as the over-detection of delirium is thought to be more beneficial when juxtaposed with under-detection [14, 15].

The routine screening for delirium is recommended for all older people seen in clinical practice. As supported by the National Institute for Health and Care Excellence (NICE), the first step in the clinical assessment conducted in acute medical setting is to rule out delirium (i.e., “think delirium”), which is an essential step prior to conducting further clinical assessments in relation to cognition and/or genuine mental health problems [16]. To date, several brief bedside instruments, most requiring professional training, have been developed to improve delirium recognition though only a handful have been properly evaluated for use in clinical setting [17]. Furthermore, there has been an emphasis on evaluation in hospital rather than community residential settings.

The two most widely recommended delirium screening tools are the Confusion Assessment Method (CAM; [18]) and the 4AT [19]. Both tools take little time to complete and are easy to administer in clinical practice.

The CAM diagnostic algorithm consists of four components: (1) acute onset of mental status changes of fluctuating course, (2) inattention, (3) disorganized thinking, and (4) an altered level of consciousness. The diagnosis of delirium is based on the presence of both components (1) and (2) and either (3) or (4). Based on the assessment, a patient is evaluated as CAM “positive” or “negative.”

A recent review, based on evaluation of 22 studies, and a total of 2442 patients, reported high sensitivity [82 % (95 % CI: 69–91 %)] and specificity [99 % (95 % CI: 87–100 %)] for CAM, with similar values for CAM-ICU, e.g., sensitivity 81 % (95 % CI: 57–93 %) and 98 % (95 % CI: 86–100 %) specificity [20], suggesting that even in trained hands, the diagnosis of delirium can be missed in up to one in every five cases if only using unimodal assessment in the form of the CAM or CAM-ICU. The relatively low sensitivity of both CAM and CAM-ICU was further highlighted by a more recent study that reported CAM sensitivity was as low as 27 % when used in the context of postoperative delirium in older people (>70 years), further arguing that the use of the CAM needs to be supported by detailed clinical assessment, based on the established delirium criteria (e.g., DSM-IV; [21]). In addition, the use of both CAM and CAM-ICU requires adequate professional training as supported by one study that found a decrease in sensitivity of 40 % when utilized by untrained staff [22]. Lastly, studies suggest that there exists variation in the application of CAM with false negatives of delirium as high as 33 % when administered by nursing staff at the bedside [23].

The 4AT had a sensitivity of 89.7 % and specificity 84.1 % for delirium against the DSM-IV-TR criteria [19], whereas a recent study conducted in patients with an acute stroke found the 4AT test to have the highest sensitivity (100 %) and a reasonable specificity (82 %) for detecting delirium in comparison to other delirium tests, including the CAM [24]. As the test also includes the Abbreviated Mental Test and Month Backwards, which are both validated short tests for cognitive impairment, it is not surprising that the 4AT also has a reasonable sensitivity (86 %) and specificity (78 %) to detect cognitive impairment [24]. Use of the 4AT as a screening tool seems promising; however, further work is necessary to evaluate the usability of 4AT when undertaken by nonmedical (i.e., nursing) professionals.

A systematic review identified several delirium screening tools for use in critically ill patients: Delirium Detection Score, Cognitive Test for Delirium (CTD), Memorial Delirium Assessment Scale (MDAS, discussed above), Intensive Care Delirium Screening Checklist (ICDSC), Delirium Rating Scale-Revised-98, and Neelon and Champagne (NEECHAM) Confusion Scale [39]. Among the latter, the Intensive Care Delirium Screening Checklist has been recommended by the Society of Critical Care Medicine [40] for delirium screening with a sensitivity of 99 % and specificity of 64 % [41].

Overall, there is a paucity of literature assessing the reliability, validity, and diagnostic test accuracy of instruments to assess people with cognitive impairment who also have an acute confusion. Although delirium superimposed on dementia (DSD) is rather common (with up to 68 % of people with dementia admitted to acute medical facilities having delirium [42]), and associated with worse outcomes (reviewed in [43]), there are still no specific diagnostic tests to detect DSD or differentiate between delirium and dementia in older people. Interestingly, delirium specialists appear to be rather confident in their clinical ability to detect delirium DSD, with one third stating that it is always easy to differentiate between delirium and dementia and 41 % being confident to discriminate between delirium and behavioral and psychological symptoms in dementia (BPSD) [43]. However, the severity of dementia, and especially Dementia with Lewy bodies, appears to be challenging when assessing for delirium [43].

A recent review on delirium screening tools used in subjects with DSD addressed the use of four delirium screening tests [CAM, CAM-ICU, Cognitive Test for Delirium (CTD), Delirium Rating Scale (DRS)], electroencephalography (EEG), and the Short Portable Mental Status Questionnaire (SPMSQ) [44]. CAM showed the highest sensitivity and specificity in this population (specificity 96–100 % and sensitivity 77 %), whereas EEG had somewhat lower sensitivity and specificity in a homogeneous patient population (100 % dementia subjects, with 91 % specificity and 67 % sensitivity) [44].

In the Richardson et al. survey [43], the delirium specialists also appear to rely on the above screening delirium tools to assess DSD in clinical practice. In addition, they also identified several useful clinical features to assess the DSD diagnosis: attention (71 %) and fluctuation in cognitive status (65 %) were the most frequent clinical symptoms identified to differentiate delirium from dementia, followed by arousability (41 %) [43]. Not surprisingly, the cognitive status was best assessed via collateral information (82 %) and medical records (59 %), whereas the more formal cognitive tests, such as the IQCODE and the cognitive rating scale, were less used in routine clinical practice [43].

Among other delirium tools that have been used for DSD screening are MDAS [25] and NEECHAM [45]. The MDAS discriminates well not only between subjects with and without delirium but also between subjects with delirium and cognitive impairment [25]. The scale is also useful in differentiating subjects with DSD and those with delirium only, with the latter subjects having significantly lower MDAS scores compared to those with delirium superimposed on dementia [46]. Since the scale is based on observational rating, it has been successfully used in patients with terminal delirium [47], as well as those in palliative [48] and critical [49] care settings.

Similarly, the NEECHAM scale [45] has also been used in delirium screening in people with dementia. This scale comprises of nine items to detect the presence and severity of acute confusion in hospitalized older adults. It also assesses cognitive processing, behavior, and physiological control and is strongly correlated with the MMSE [45]. The assessment is fairly brief and can be completed by nursing staff while assessing the vital signs.

Following basic principles of history and examination, diagnosis of delirium is often aided by a range of clinical and laboratory-based investigations (Table 7.5). The most common investigations involve blood and urine tests aimed at determining underlying organic causes and risk factors of delirium. Radiological and functional studies are also occasionally used in conjunction with basic investigations. Expert interpretation is required when it comes to the context of organic pathology as a cause of delirium. Frequently, investigations may be reported as abnormal or incidental with a limited or mixed role when it comes to an immediate cause of delirium. For example, altered (slow-wave) electroencephalography (EEG) activity can be present in both delirium and dementia (e.g., Alzheimer’s disease) and thus may necessitate reference with any preceding records. Conversely, structural changes of cerebral white matter hyperintensities and general and medial temporal lobe atrophy are thought to be usually associated with dementia and unrelated to delirium incidence or severity [51]. Examples of investigations used as a diagnostic adjunct are listed in Table 7.5. 123I-FP-CIT SPECT imaging can also be helpful in differentiating delirium from other causes of neuropsychiatric symptoms (e.g., dementia with Lewy bodies, Parkinson’s disease with and without dementia, and atypical causes of parkinsonism such as corticobasal degeneration and multiple system atrophy) though abnormal scans need to be interpreted in the context of other potential confounders such as cerebrovascular disease. More recently, animal studies hint at the potential competitive inhibition of certain medications (e.g., benzodiazepines, antipsychotics) with the presynaptic binding of the FP-CIT tracer though the clinical significance of this remains uncertain [52].

As yet, detailed neuropathological studies on delirium are lacking [53–55], and consequently the current neurobiological understanding of the pathogenesis of delirium is based largely on correlative clinico-biochemical studies. The pathophysiological link between delirium and a broad array of precipitating factors remains difficult to establish as most of these conditions occur without identifiable involvement of the brain (Table 7.5). Despite the uncertainty on how different etiological causes may evoke similar symptoms, it is thought that a common feature in delirium is the acute and transient impairment in homeostatic balance of the central nervous system (CNS) comparable to the concepts of renal or hepatic insufficiency. There is clinical and neuropathological evidence suggesting that preexisting dementia significantly increases susceptibility of the CNS to deleterious effects of acute insults [56]. Synaptic disconnection may be a major contributor to delirium risk, together with other aspects of dementia, such as microglial activation and chronic hypocholinergic function.

Accumulating evidence shows that acute peripheral inflammatory stimulation [e.g., with IL-1 or lipopolysaccharide (LPS)-induced inflammatory response] induces a cascade of functional and structural changes in the CNS with consequent neurochemical and functional disturbances in different brain structures [57]. In animal models, these changes underlie acute and transient disturbances in cognition and behavior (the so-called ‘sickness behavior syndrome’) supporting that an acute neuroinflammatory response is likely to be implicated in the pathophysiology of delirium (Table 7.6). Elevated levels of plasma C-reactive protein (CRP) have been reported in some [58, 64, 65], but not all [61–63], studies that involved medical and surgical patients with delirium. Although one study reported that subjects who developed delirium following elective hip arthroplasty had a higher ratio of proinflammatory to anti-inflammatory cytokines [9], higher plasma levels of IL-6, IL-8, and IL-10 in patients with delirium have not been consistently replicated. In a case–control study, only IL-8 (but not IL-1b, IL-6, IL-10, IL-12p70, and TNF-α) was found to be increased in cerebrospinal fluid (CSF) of patients with preoperative or postoperative delirium [77]. Preoperative CSF levels of interleukin-1 receptor antagonist and interleukin-6 were significantly lower in elderly hip fracture patients who developed delirium postoperatively [73]. Brains of patients who developed delirium near death also had higher microglial and astrocytic activation coupled with increased levels of IL-6 compared with age-matched controls without delirium postmortem [54].

The pathophysiology of delirium can also involve a dysfunctional centralized response to acute infection or injury through exaggerated activation of efferent networks including the hypothalamic–pituitary–adrenal (HPA) axis. In fact, cortisol levels have been found to be elevated in plasma and cerebrospinal fluid of patients with delirium in various medical and surgical conditions which is consistent with the hypothesis of an “aberrant stress response” associated with delirium [96, 97]. However, it remains unclear if cortisol elevation is the result of a primary dysfunction of the HPA axis or represents a secondary change to a more intense inflammatory reaction.

So far, two proteomic studies have been published in delirium subjects. One study utilized urine from subjects following cardiac surgery but found no distinct protein fingerprint specific to delirium [98]. The other study reported a dysregulated protein expression in the CSF of 17 delirious patients compared to controls with mild Alzheimer’s disease. These proteins included apolipoproteins, chromogranin/secretogranins (downregulated in most delirium subjects), and inflammation-related proteins (mostly upregulated in delirium) [99].

It remains largely unknown how systemic peripheral changes (i.e., inflammation, dehydration, electrolyte imbalance) translate to acute cognitive and behavioral symptoms. Cortical atrophy, ventricular enlargement, and increased white matter lesions may predispose individuals to develop delirium [100].

Inattention, a core feature of delirium, is thought to result from dysfunctional communication between prefrontal and parietal cortices and/or impaired neuromodulation from cholinergic basal forebrain neurons [101]. A study using resting-state functional MRI during an episode of delirium showed disruption in reciprocity of the dorsolateral prefrontal cortex with the posterior cingulate cortex and reversible reduction of functional connectivity of subcortical regions [102]. Adequate function of attentional networks requires sufficient arousal provided by the ascending reticular activating system which is located in the upper brain stem tegmentum and central thalamus [103]. Connections between these subcortical structures are associated with acetylcholine and dopamine. The role of cholinergic deficiencies has received the greatest amount of attention, as this neurotransmitter system is involved in sleep, attention, arousal, and memory. Dopamine excess may also be involved as it exerts a regulatory influence over the release of acetylcholine [102].

In the absence of human studies, studies in aged animals on inflammation provide accumulating evidence that supports an acute systemic inflammatory process (e.g., LPS challenge) that results in acute hippocampal dysfunction and cognitive inflexibility [104], as well as deficits in attention/executive function [101]. These findings support a direct link between inflammation and the defining clinical symptoms of delirium.

Studies investigating the association of genetic polymorphisms with delirium have provided conflicting results. The epsilon 4 allele of apolipoprotein E (ApoE ε4) was a risk factor for postoperative delirium in patients undergoing noncardiac surgery [105] and a predictor of longer duration of delirium in critical patients [106]. However, this was not confirmed in other studies that included medical [107] and surgical [108, 109] patients. Thus, the link between the APOE genotype and delirium remains to be further elucidated.

Variations in the SLC6A3 gene and possibly the DRD2 gene [110], but not catechol-O-methyltransferase gene [111], and IL-6 and IL-8 gene polymorphisms [112] have been associated with delirium in patients after hip fracture. This suggests that genetic and/or phenotypic expression of these cytokines does not play a role with the actual physiological inflammatory response associated with delirium though further study is required to assess for potential genetic associations of other cytokines. It is worth highlighting that the COMT val158met polymorphisms are associated with impaired executive function in Parkinson’s disease (PD) [113] and as such may have a role in attention. Since PD subjects are at higher risk of developing delirium [114], it is interesting to speculate that this genetic polymorphism may contribute to the susceptibility of people with PD to delirium. In subjects undergoing coronary artery bypass surgery, the presence of AG haplotype of GRIN3A gene (a genetic variation of NR3A subunit of NMDA receptor) independently increased the risk of postoperative delirium, whereas GRIN2B and 5HT2A gene polymorphisms were not associated with delirium [115]. None of the polymorphisms of MTNR1B gene were found to be associated with the occurrence of delirium [116].

Numerous challenges exist with distinguishing delirium and dementia. Generically, older people with multiple comorbidities and overall poor medical health can present with an impaired level of consciousness and/or communication which prevents the use of tools that rely on patient collaboration, e.g., cognitive assessment tools, such as the MMSE [31]. Furthermore, disorientation and memory problems are core characteristics of cognitive impairment. These symptoms cannot always be attributed as a result of delirium alone, and presence of an alternative cognitive impairment (which may or may not be diagnosed) should be considered, for example, the presence of comorbid dementia. Despite this, it is known that the symptoms of inattention and disturbance of consciousness may be more sensitive and are time efficient for the diagnosis of delirium despite the potential weakness in sensitivity [117]. Hence, it is not surprising that these two symptoms were incorporated in the CAM.

The above vignette is an example of hypoactive delirium. It illustrates the importance of obtaining good collateral information especially on pre- and co-morbid cognitive and behavioral function in an elderly individual with suspected delirium. This helps differentiate between delirium and dementia, as well as functional mental health illnesses. The British Geriatric Society (BGS, 2006; [118]) provided an algorithm to aid the diagnosis between dementia and delirium. This algorithm includes the use of an informant questionnaire (IQCODE-SF) and cognitive assessment (clock drawing test, MMSE). One study found that lower MMSE scores (<24/30) in particular were associated with a 5.5-fold increased risk of developing delirium [119] and, as such, can be used to predict at-risk subjects.

Understandably, it is not uncommon that individuals with delirium can experience difficulty completing cognitive assessments especially due to an impaired level of consciousness (coma or stupor), inability to communicate, or rapidly deteriorating medical condition [117]. This suggests cognitive tests may not be an appropriate instrument and its weighting in specific algorithms (diagnostic pathways) may need to be reconsidered especially in the elderly who present with fluctuating confusion with the setting of an acute medical illness. To address this, the BGS algorithm can be easily modified to include an adequate delirium screening tool such as the CAM or 4AT test in combination with the IQCODE-SF and/or structured collateral information (Fig. 7.1).

Case Vignette 2 underlines a number of important principles. Firstly, it is important to exclude any other organic cause(s) of this presentation. Secondly, iatrogenic causes of clinical presentations akin to delirium may result from inception or discontinuation of a medication. In this case, it is known that dopamine agonists can cause side effects especially at higher dosage regimes. Common side effects include impulse control disorders, visual hallucinations, and excessive daytime somnolence. More recently, the phenomenon of dopamine agonist withdrawal syndrome is being increasingly reported and thought to resemble other psychostimulant withdrawal syndromes that lack a therapeutic response to levodopa, antidepressants, and anxiolytics [120]. In terms of treatment, the use of clozapine, an atypical antipsychotic, is used in select cases of PD-associated psychosis.

Further principles outlined include the need to monitor for treatment response and adverse events as well as ongoing follow-up to ensure resolution. There is more favorable evidence for the use of clozapine over other agents such as quetiapine; however, this does not come without its drawbacks as routine monitoring is required to ensure potentially deleterious adverse effects such as agranulocytosis do not occur. The ongoing care and follow-up of the patient is important, and the risk of further cognitive decline (or progression to Parkinson’s disease dementia) is increased if ongoing cognitive issues persist especially in the setting of complex visual hallucinations.