Delirium superimposed on dementia (DSD)—a practical approach

Delirium is common in people living with dementia. Delirium superimposed on dementia (DSD) should be considered whenever the person is different from usual in their engagement, function, distress or behaviour. Differentiating DSD from dementia alone can be challenging, particularly in advanced dementia, but existing tools can help. Good dementia care, delirium prevention and delirium management overlap, which helps implementation. Delirium prevention/management needs to be adapted to the person’s cognitive abilities and augmented at times of extra risk. Delirium is an acute or subacute neuropsychiatric disorder, with multiple somatic causes, including infection and metabolic disturbances. People with dementia are at high risk for delirium superimposed on dementia (DSD). In the Cork delirium-dementia studies, 72%–73% of delirium in older adults was DSD [1, 2]. Equally, almost 50% of older hospitalised adults with dementia have DSD [3]. DSD is linked to adverse outcomes, compared to dementia alone, like longer hospitalisation [3], mortality [3, 4], functional decline [3] and need for residential care [3]. DSD also has a worse mortality rate than delirium alone [4]. Delirium accelerates cognitive decline [5]. Delirium is also highly distressing for patients, caregivers and staff and efforts should be made to prevent and limit it in all populations. Dementia pathology, neuroinflammation and neurotransmitter imbalances, including acetylcholine deficiency, are likely significant predisposing risks for delirium [6], along with psychotropic polypharmacy, which is common in dementia. In a surgical cohort, unfavourable amyloid and tau profiles, but not Apoe E status (without dementia), conferred a higher delirium risk [7]. Common dementia-related events which could precipitate delirium include fall-related injuries, pneumonia and pressure ulcers. Reported DSD risks include delayed hip fracture surgery, malnutrition, acute infection, psychotropic medications and male gender, and possibly dementia severity and age [3]. Some dementia types may carry greater DSD risk. Lewy body dementia (LBD) (i.e. dementia with Lewy bodies or Parkinson’s disease dementia) causes impaired visual attention shifting, prominent visual hallucinations [8], cognitive fluctuations, severe daytime sleepiness and visuospatial issues. LBD has very reduced acetylcholine levels, even in early stages [8], and dopamine deficiency-hypersensitivity, which likely lowers the delirium threshold, further exacerbated by dopaminergic and anticholinergic medications. Thus, a person with LBD is ‘delirium-primed’, while paradoxically DSD is more difficult to diagnose. Similarly, vascular dementia has an increased delirium risk [9], while features like impaired attention, alertness and emotional lability, and acute onset or step-wise progression, can cause diagnostic challenges for DSD. However, many people with Alzheimer’s disease have concomitant cerebrovascular disease and people with PD often have Alzheimer pathology, so that dementia phenotype may not fully reflect delirium risk. DSD is frequently missed; symptoms can be mistakenly interpreted as non-cognitive symptoms of dementia (NCSD). Although non-pharmacological approaches to both conditions overlap, diagnosing DSD is important for its prognostic implications, as above, and prompts a search for reversible causes. Thus, delirium consideration is the first step of NCSD management [10]. However, assessment is challenged by dementia-related impairments, where worsening cognitive and communication impairment limit a person’s ability to express delirium symptoms or answer test questions. Instead, a person may be more restless and agitated, or, conversely, more subdued. Collateral history from an informant who knows the person with dementia well is thus crucial in delirium detection, as they will notice such a change from normal. The Single Question in Delirium, asking if the person is ‘more confused lately’, should retain value in advancing dementia, when posed to a reliable informant, if ‘confused’ is interpreted as cognitively or behaviourally different to normal. Intuitively, tools which assess changes in alertness, fluctuation and inattention have value. As delirium typically alters arousal levels, assessing alertness is key in detecting DSD. Supporting this, a large multicentre study found that deviations from normal/modified Richmond Agitation and Sedation Scale scores effectively distinguished DSD, with good sensitivity and specificity [11]. Among delirium detection tools, the 4AT is weighted for altered alertness and acute change/fluctuation (4 marks each), with either being sufficient to indicate delirium; the remaining attention and orientation items (4 marks combined) may have floor effects. A sub-analysis reported 94% sensitivity and 65% specificity in dementia [12]; in ED patients with dementia, sensitivity was 92% and specificity was 79% [2]. The 3D-CAM has reported good sensitivity and specificity in dementia, as has the 3-item screener [13] (reciting weekdays backwards; orientation to place; drowsiness) and combined arousal and attention tests [14]. The RADAR tool [15] is worth considering. This avoids any cognitive testing, by assessing difficulty following instructions, drowsiness and slowed movements, during a medication round, thus incorporating a quick functional assessment in a routine task. Finally, the 4-DSD [alertness; altered function; attention; acute change or fluctuation] was created specifically to detect DSD, with good sensitivity and specificity even in advanced dementia (82% and 56%, respectively) [16]. Beyond formal tools, as long-term memory is relatively preserved in earlier dementia, conversations about important events from the person’s life might reveal attention, arousal or memory deficits that suggest DSD. In advanced dementia, assessing alertness and engagement (e.g. following movements with gaze, smiling appropriately, mirroring actions) may be more useful. There is no reason to believe that DSD would have less response than delirium alone to non-pharmacological interventions. A systematic review of prevention interventions found low to moderate quality evidence for reduced delirium incidence, with the strongest evidence from a small RCT in a hip fracture cohort (n = 64) [17]. Good dementia care and delirium prevention/management overlap significantly, aiding feasibility. Many delirium prevention/management components are easily implementable for people with dementia, like medication review, facilitating family support, prioritising undisturbed sleep and avoiding constipation or unnecessary catheterisation. Some may need extra support, such as using hearing or vision aids and maintaining mobility. Where clocks/calendars can no longer be understood, it may be helpful to augment day/night differentiation through noise and lighting controls and different day/night clothing, with temporal cues via meals and typical morning/evening activities. In terms of pharmacological treatment, there is little evidence for effectiveness in delirium management in general. Moreover, a person with dementia is at high risk of psychotropic side effects, such as sedation and falls from benzodiazepines. People with LBD are at particular risk from antipsychotics, including death, and their use should be avoided [10]. In conclusion, delirium assessment is a priority in a person with dementia who is deteriorated or ‘different’. Existing tools may help to a degree, with a focus on arousal, fluctuations and lowered attention/engagement. Skilled clinical assessment, with appropriate consideration of dementia-related impairments, is required to confirm DSD and identify potentially reversible causes. Detecting DSD is not an academic exercise. It triggers a search for potentially reversible causes, enhanced care for the person, and communication with the family in terms of explanation for the person’s presentation and alerting them to the poorer prognosis. As with all delirium, prevention is key. None. None.