Authors: Mfon E. Umoh (1Division of Geriatric Medicine and Gerontology, Department of Medicine, Johns Hopkins School of Medicine, Baltimore, Maryland), Dennis Fitzgerald (2Department of Neurology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts), Sarinnapha M. Vasunilashorn (3Division of General Medicine, Department of Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts; 4Department of Epidemiology, Harvard T. H. Chan School of Public Health, Boston, Massachusetts), Esther S. Oh (1Division of Geriatric Medicine and Gerontology, Department of Medicine, Johns Hopkins School of Medicine, Baltimore, Maryland), Tamara G. Fong (2Department of Neurology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts; 5Aging Brain Center, Marcus Institute for Aging Research, Hebrew SeniorLife, Boston, Massachusetts)
Categories: Article, delirium, dementia, delirium superimposed on dementia, cognitive decline, geriatrics
Source: Seminars in neurology
Authors: Mfon E. Umoh, Dennis Fitzgerald, Sarinnapha M. Vasunilashorn, Esther S. Oh, Tamara G. Fong
Delirium and dementia are common causes of cognitive impairment in older adults. They are distinct but interrelated. Delirium, an acute confusional state, has been linked to the chronic and progressive loss of cognitive ability seen in dementia. Individuals with dementia are at higher risk for delirium, and delirium itself is a risk factor for incident dementia. Additionally, delirium in individuals with dementia can hasten underlying cognitive decline. In this review, we summarize recent literature linking these conditions, including epidemiological, clinicopathological, neuroimaging, biomarker, and experimental evidence supporting the intersection between these conditions. Strategies for evaluation and diagnosis that focus on distinguishing delirium from dementia in clinical settings and recommendations for delirium prevention interventions for patients with dementia are presented. We also discuss studies that provide evidence that delirium may be a modifiable risk factor for dementia and consider the impact of delirium prevention interventions on long-term outcomes.
Delirium and dementia are the most common causes of cognitive impairment in older adults. Delirium, an acute state of confusion characterized by impairments in attention and cognition, has been linked to the chronic and progressive loss of cognitive ability as seen in dementia. Individuals with dementia are at higher risk for delirium, and delirium itself is a risk factor for incident dementia.^1^ Delirium is independently associated with a decrease in cognitive measures months after resolution,^2^ and there is overwhelming evidence that delirium is an independent predictor of adverse outcomes and faster cognitive decline.^3,4^ When a person with dementia (PWD) develops delirium, the clinical syndrome is referred to as delirium superimposed on dementia (DSD). The pathogenesis underlying the intersection between these conditions remains an active area of ongoing research.
Delirium is prevalent in hospitalized patients and is associated with significant morbidity and mortality.^5^ In the United States, over 2.6 million adults aged 65 years and above develop delirium each year.^6^ Delirium occurrence has been estimated at 23% among adult medical patients in secondary care.^7^ Cumulative costs attributable to delirium are soaring,^8^ with studies estimating the burden of delirium on the health care system ranging between 152 billion annually.^9^ The prevalence of dementia, for which Alzheimer’s disease (AD) is the most common form, is estimated at 57.4 million cases globally, with the prevalence of AD and related dementias (ADRDs) projected to triple to more than 150 million people by 2050.^10^ Dementia is the seventh leading cause of death and one of the major causes of disability and dependence in older adults globally.^11^ PWD have higher utilization of healthcare services and higher healthcare costs than those without dementia in the United States.^12,13^ The impact of delirium and dementia on older adults is notable.
This review will describe the interrelationship between delirium and dementia, two highly prevalent and important conditions. Strategies for evaluation and diagnosis that focus on distinguishing delirium from dementia in clinical settings will be discussed. Current epidemiological, clinicopathological, neuroimaging, biomarker, and experimental evidence that links delirium and dementia are presented. Importantly, we highlight studies reviewing the effect that delirium has on long-term outcomes in patients with dementia, and we discuss the potential role of delirium prevention in patients with dementia. Lastly, we consider the broader impact of delirium beyond acute hospitalization.
While dementia and delirium are common causes of cognitive impairment among older adults, they are distinct conditions. Indeed, dementia should not be diagnosed in the face of delirium, and the diagnosis of delirium should not be made when symptoms can be “better accounted for by a preexisting, established, or evolving dementia.”^14^ Clinical diagnoses can be difficult even for experienced clinicians, and can be especially challenging in PWD at more advanced stages. Delirium symptoms can persist for weeks or even years,^13–18^ and the recognized conditions of “persistent delirium” and “reversible dementia” challenge the extent by which delirium and dementia are truly discrete conditions.^15^
Features differentiating delirium and dementia can be subtle. For example, in older adults with known or possible dementia, drowsiness or lethargy from delirium may be missed or mistakenly attributed to the underlying dementia. There are some signs and symptoms that can be helpful (Table 1).^16–18^ The onset of delirium is abrupt over hours to days; the observation of an acute change in cognition and alertness may be the most reliable means for identifying delirium, but it requires knowledge of the individual’s baseline cognitive status. Attention and level of consciousness are typically reduced and fluctuate in delirium, whereas these domains typically remain intact until the advanced stages of dementia. As an exception, in Lewy body dementia, fluctuation in symptoms is a key defining characteristic.^19^ Hallucinations in delirium are often visual, tactile, and poorly formed, and differ from hallucinations in AD which may be auditory or visual and are typically more distinct.^20,21^ Thought abnormalities in delirium are often “flight of ideas” characterized by disconnected or incoherent thoughts, in contrast to AD, where patients have paranoid, non-bizarre, and simple delusions.^21^ Language impairment in delirium can include irrelevant or tangential speech, reduced phrase length, and comprehension deficits,^22^ while patients with moderate to severe AD may develop aphasia, with very little to say both spontaneously and in conversation. Language impairments can create a significant challenge with using certain diagnostic scores for delirium. Sleep–wake cycle is frequently reversed in both delirium and AD. Ultimately, the differentiation may depend on the presence of an acute change in mental status or behavior from baseline noted by an informed caregiver, or may be established only in retrospect by resolution of symptoms. In the face of uncertainty, acute mental status changes should be treated as delirium until proven otherwise.
The diagnosis of delirium follows the standard clinical approach.^6^ For PWD, a major challenge in the diagnosis of DSD is the lack of validated delirium instruments appropriate for this patient population.^23^ A prior systematic review of delirium instruments used in studies that included PWD found that while some instruments showed promise, many of the studies were limited by small sample sizes and did not specify dementia subtype or severity.^24^ A few instruments have been specifically designed for, or tested in, persons with DSD, including the Six Item Cognitive Impairment Test (6-CIT),^25^ 4 A’s Test (4AT),^26^ the 4-item delirium superimposed on dementia tool (4-DSD),^27^ the 3-D Confusion Assessment Method (3D-CAM),^28^ and the Ultra-brief CAM (UB-CAM)^29^ (Table 2). Additionally, the Richmond Agitation and Sedation Scale has been studied and found to have moderate sensitivity and very high specificity for the detection of DSD.^30^ Another study using a combined arousal-attention assessment correctly classified 93% of patients in the dementia subgroup with delirium, with a sensitivity of 94% and specificity of 92% (AUROC: 0.98).^31^ These tools help establish symptomatology in patients and can be very useful clinically to distinguish these interconnected conditions. Early and reliable detection of delirium and DSD is of critical importance, as the evaluation and clinical management of delirium, dementia, and DSD are distinct.
Cognitive impairment and dementia are substantial risk factors for delirium. Table 3 summarizes studies that have examined preexisting cognitive impairment or dementia as risk factors for delirium, adjusted for important confounding variables.^32–50^ The studies include mean ages ranging from 59 to 85 years, recruited from diverse hospital settings, including medical or geriatric medicine wards, emergency department (ED), and surgical services. Cognitive baseline status was determined by a variety of approaches, including brief cognitive screening tests, proxy-based measures, clinician diagnosis, or chart documentation of dementia. Delirium was also measured by a variety of approaches, including delirium instruments, clinical criteria, and chart review. The rate of delirium ranged from 9 to 44% across these studies. Across studies, baseline cognitive impairment or dementia was a substantial independent risk factor for delirium, consistently increasing delirium risk by 2- to 10-fold in some populations (Table 3). Among cardiac surgery patients, a recent systematic review and meta-analysis found baseline cognitive impairment to be associated with an eightfold increased risk of delirium.^51^ whereas a systematic review and meta-analysis of 13 studies with 3,183 patients with stroke found that patients who had baseline cognitive decline/dementia had a mean 3.70-fold higher risk of delirium than those without. Interestingly, the analysis of the stroke population also found that participants in studies published between 2010 and 2020 had a higher delirium risk than those in studies published between 1990 and 2009,^52^ possibly due to improved delirium recognition.
Not only is dementia a risk factor for delirium occurrence, but delirium itself is an independent risk factor for long-term cognitive decline and dementia. This has been shown in several cohort studies, both prospective and retrospective analyses of memory clinic outpatients, hospitalized ICU inpatients, and patients undergoing elective surgery (Table 4). Findings from these studies consistently indicate that an episode of delirium can accelerate the rate of cognitive decline and increase the risk of incident dementia. Cognitive outcomes were determined using a variety of measures, including full neuropsychological assessments, screening of global cognitive measures, clinician diagnosis, or consensus panel diagnosis. Despite the varied methodologies across these studies, the findings are consistent in demonstrating significant associations between delirium and cognitive decline. Among these studies, delirium was significantly associated with cognitive impairment 3 months or longer after the delirium episode. A recently published meta-analysis included 24 studies totaling 3,562 patients who experienced delirium and 6,987 controls who did not found a significant association between delirium and long-term cognitive decline, with an estimated effect size (Hedges’ g) for 23 studies of 0.45 (95% CI, 0.34–0.57; p < 0.001), which is an effect size large enough to be associated with clinically significant differences.^53^ Another study using a competing risk regression model to determine the sub-hazard ratio (SHR) for each exposure with death as a competing risk found that delirium was associated with an increased dementia risk (SHR = 8.70, 95% CI: 3.26–23.24, p < 0.001).^54^
There are few large studies with long-term follow-up of delirium in subjects free of dementia at baseline. One retrospective study followed up patients without a dementia diagnosis from their first episode of delirium for ~2 years. The investigators found that 27% of the study cohort had a subsequent diagnosis of dementia, and 45% died without a diagnosis of dementia.^55^ In a study of 560 adults, aged 70 years or more without dementia scheduled for major surgery, patients with postoperative delirium (POD) had significantly lower preoperative cognitive performance, greater immediate (1 month) impairment, equivalent recovery at 2 months, and significantly greater long-term cognitive decline compared with the nondelirium group.^56^
PWD who develop delirium have worse outcomes compared with those with dementia without an episode of delirium. DSD is associated with poorer clinical outcomes compared with dementia alone, including an accelerated trajectory of cognitive decline, greater functional impairment, prolonged hospital stays, higher rates of rehospitalization, higher cost of care, and increased mortality.^57–63^ A recent systematic review identified that persons with DSD compared with PWD alone had an average length of stay that was 3 days longer.^63^ Additionally, individuals with cognitive dysfunction across multiple domains were at a greater risk for institutionalization at the time of hospital discharge and greater cognitive decline at 1 year postdischarge compared with persons with dementia alone.^63^
The impact of delirium goes beyond the hospital setting. Studies evaluating delirium beyond the hospital setting are less common, as most studies evaluate delirium within a postoperative context, ICU setting, or within medical wards. In a 24-month study of a prospective cohort of 139 older adults with dementia which followed patients daily during hospitalization and 1 month post-hospitalization, those with delirium had longer hospital lengths of stay, poorer function at discharge, greater cognitive decline, and a 25% short-term mortality rate compared with 9.5% in those with no delirium.^64^ Another prospective cohort found that of 5,570 individuals admitted from a private residence who survived to discharge, those with DSD were less likely to be living at home (odds ratio [OR]: 0.25), compared with those with dementia (OR: 0.43), unspecified cognitive impairment (OR: 0.55), and delirium alone (OR: 0.57) at the time of follow-up.^65^ The effect of delirium on cognitive function at 1 year, both its duration and number of episodes, in patients with and without dementia was associated with cognitive decline (−1.8 mini-mental state examination points [95% CI: −3.5 to −0.2]), an increased risk of new dementia diagnosis at follow-up (OR: 8.8 [95% CI: 1.9–41.4]), and more than one episode and more days with delirium (>5 days) were associated with worse cognitive outcomes.^66^ A systematic search and literature review on delirium in primary care and institutionalized long-term care (LTC) found that general practitioners report data that show a very low prevalence of delirium and are likely missing many cases; the studies included found that, in these populations, age and cognitive decline significantly increased the risk of delirium.^67^
To compare long-term outcomes among newly admitted skilled nursing facility (SNF) patients with delirium, incident ADRD, and both conditions, a retrospective cohort study of Medicare beneficiaries who entered an SNF from a hospitalization with a minimum 14-day stay (n = 100,832) from 2015 to 2016 found that a positive delirium screen alone increased the risk of transfer to LTC and increased the risk of death in the first 100 days after admission regardless of ADRD diagnosis.^68^ Another study, a nationwide retrospective cohort study examining the effect of early detection of delirium in SNFs that included 1,175,550 Medicare enrollees who entered the SNF from a hospital and had no prior diagnosis of dementia found positive screening of delirium in 7.7% of cases, most occurring within the first 7 days of SNF admission (62.5%).^69^ Results from that study suggest that among older adults not previously diagnosed with dementia, a positive screen for delirium was significantly associated with a higher risk of ADRD diagnosis after admission to an SNF, with the highest risk in the first days of the patient’s stay and in those with the least cognitive impairment, suggesting the ADRD was overdiagnosed in the setting of delirium.^69^
Few studies exist evaluating delirium outcomes in the outpatient setting. One retrospective cohort study of 109 older patients with delirium referred to a memory clinic found a delirium prevalence of 3.6%, and that delirium was associated with worse functional (ADL 2.95 ± 1.95 vs. 2.16 ± 1.84) and cognitive (MMSE 13.88 ± 8.96 vs. 11.0 ± 9.49) status after 6 months compared with baseline.^70^ In this study, the mortality rate was 29.4%. Of the 28.3% of the cohort admitted to an LTC facility after the episode of delirium, more than half were hospitalized during the follow-up (of the 109 patients with delirium, 85 were managed at home and 24 were hospitalized); patients who were hospitalized had more severe behavioral symptoms during the delirium episode. In this study, there was no difference in mortality and institutionalization according to the home or hospital management.^70^
The overlap between delirium and dementia has been suggested to stem from shared pathophysiology, including acceleration of dementia pathology, failure of the vulnerable brain to show resilience in response to an acute stressor (decreased cognitive reserve), neurotoxicity, and several other pathways (Fig. 1). Several studies have tried to understand the neuropathological correlates of these conditions.
One retrospective population-based study, Vantaa 85+, examined the impact of delirium in a cohort of 553 individuals aged 85 years and older and neuropathological markers of dementia.^71^ The cohort was followed up for incident dementia and other markers of health status for up to 10 years. They found relationships between dementia and measures of neurofibrillary tau, amyloid burden, apolipoprotein E (ApoE) ε4, vascular lesions, and Lewy body pathology were strongest in the absence of a history of delirium. While the study was not powered to be conclusive, the results suggest that when delirium is part of the dementia trajectory, the pathological substrates may be different from conventional dementia pathology, such as AD, vascular or Lewy body pathology, suggesting the acceleration of cognitive decline following delirium might result from an alternative pathway leading to neuronal damage. Another study using human tissue and mouse models of delirium found that for every MMSE point lost, the risk of incident delirium increased by 5% (p = 0.02.).^38^ Another study which included neuropathology from three separate cohorts found that the change in MMSE score in the 6 years prior to death was fastest in patients with delirium and higher dementia pathologic burden, in comparison to patients with no delirium and little dementia pathologic burden, who showed the slowest decline in MMSE score.^72^ This suggests that delirium might interact with classic dementia pathology in an independent but interrelated way, leading to cognitive impairment and dementia.^72^
Despite the growing epidemiologic evidence of the association between delirium and dementia, we know far less about the pathophysiology linking the two conditions. Among emerging work on the shared biology underlying delirium and dementia across different study settings (e.g., medical, surgical, intensive care unit [ICU]), a common theme surrounding the immune response has emerged. Increasing evidence from blood and cerebrospinal fluid (CSF) points to the role of the stress response and neuronal injury.
A summary of the potential markers shared between delirium and dementia has been published.^73^ Briefly, indicators of systemic inflammation include acute phase reactant C-reactive protein (CRP),^73,74^ proinflammatory cytokines interleukin (IL)-6, and IL-8,^75,76^ and chitinase 3-like protein (YKL-40), a marker of type 2 immune response.^77^ Plasma markers of neuronal injury include neurofilament light chain (NfL) and S100 calcium-binding protein β (S100β).^76,78^ CSF markers of inflammation shared between delirium and dementia include IL-8 and total protein,^78,79^ as well as a soluble fragment of triggering receptor expressed on myeloid cells (sTREM2), a marker of neuroinflammation.^80^ A meta-analysis aimed to identify AD and AD-related (AD/ADRD) biomarkers associated with POD included 28 studies and quantified the pooled differences in some of these blood inflammatory markers (CRP and IL-6) and neuronal markers (S100B and NfL) with POD.^81^
In addition to these plasma and CSF markers, markers of brain vulnerability likely play important roles in the underlying shared pathophysiology linking delirium and dementia. These indicators of brain vulnerability include the genetic risk marker ApoE ε4,^82–84^ the genetic factor most strongly associated with AD, and other indicators associated with the risk for AD (e.g., Aβ1–40, Aβ1–42, total (t)-tau, phosphorylated (p)-tau181, p-tau217).^85–90^ Moreover, consideration of several AD biomarkers provides further opportunities to identify brain vulnerability using the ATN biomarker framework that distinguishes AD causes from non-AD causes of cognitive impairment based on β-amyloid deposition (A),^91^ pathological tau (p-tau, T), and neurodegeneration (N).^92^
Taken together, this body of work supports an emerging model (Fig. 1) for delirium pathophysiology in which individuals predisposed to a heightened inflammatory response are at increased risk for delirium when exposed to an acute stressor, such as surgery or infection.^93,94^ Under certain conditions (i.e., indicators of a vulnerable brain [e.g., the presence of genetic risk modifiers or heightened levels of AD biomarkers]), these systemic inflammatory mediators may cross the blood–brain barrier (BBB), activate brain microglia, and set up a process of neuroinflammation, which, if sustained, can cause permanent neuronal injury and result in downstream outcomes, including delirium, long-term cognitive decline, and, in some individuals, dementia. Although current empirical data are consistent with this model of inflammation, it is unlikely that this single mechanistic pathway can be implicated in all cases of delirium and dementia given the complexity of each condition, individual, and the complex interrelationship between the two.
Advances in technologies, especially multiple “omics” approaches, present the opportunity to efficiently identify additional biological pathways involved in the pathogenesis of delirium and dementia.^95^ Initial proteomics findings have confirmed the associations of commonly examined inflammatory markers (e.g., CRP and IL-6), and have identified novel markers associated with delirium and dementia (e.g., YKL-40).^77^ Overall, findings from metabolomics and lipidomics approaches have identified pathophysiological mechanisms pertaining to neuroinflammation, oxidative stress, energy metabolism, and neurotransmitter imbalances reported in studies of delirium and AD.^95–97^
Mounting evidence for the role of inflammation continues to emerge. A systematic review focused on postoperative delirium and its relationship with biomarkers for dementia included 28 studies, most focused on inflammatory and neuronal injury biomarkers.^81^ Two inflammatory biomarkers (IL-6 and CRP) showed a significant relationship with POD (IL-6 n = 10, SMD: 0.53, 95% CI: 0.36–0.70; CRP n = 14, SMD: 0.53, 95% CI: 0.33–0.74), while two neuronal injury biomarkers (blood-based S100B and NfL) were positively associated with POD (S100B n = 5, SMD: 0.40, 95% CI: 0.11–0.69; NFL n = 2, SMD: 0.93, 95% CI: 0.28–1.57 [vs. MGM1]).^81^ Another study found mechanisms of neuroinflammation involved in both delirium and dementia, specifically, as observed in a triggering receptor expressed on myeloid cells 2 (TREM2) that encodes expression of an innate immune receptor in the brain expressed by microglia.^80^ In 120 participants with and without preexisting dementia who underwent hip fracture surgery, the level of CSF-soluble fragment of TREM2 (sTREM2) was reported to be higher in prodromal and asymptomatic AD compared with participants without preexisting dementia.^80^ Higher CSF levels of sTREM2 were observed among participants with delirium who did not have preexisting dementia relative to participants with delirium and preexisting dementia, particularly among patients developing delirium after CSF sampling.^80^ This finding underscores the involvement of neuroinflammation in delirium, and suggests separate responses in patients with or without preexisting dementia. Findings in animal models further highlight the role of inflammation and neuroinflammation in delirium and dementia pathophysiology.
Overall, emerging evidence suggests that biomarkers that can inform mechanistic insights into delirium and dementia include markers of inflammation, brain injury, and cellular stress. Future areas of research aim to understand (1) underlying brain injury mechanisms triggering and/or sustaining delirium, (2) dysfunctional neural circuitry of delirium, and (3) interactions of biomarkers with clinical risk factors that may contribute to long-term downstream consequences.
Early neuroimaging investigations were largely limited to case–control or retrospective cohorts, with findings often heterogeneous. More recently, collaborations using standardized data including neuroimaging, neuropsychology testing, and fluid biomarkers collected at multiple time points have undertaken the shared aim of rigorously studying delirium and its relation to long-term outcomes. Innovative structural and functional neuroimaging applications, and advancements in statistical analysis of functional studies, have allowed for the study of associations between dementia and delirium. Evidence is summarized in Table 5.
Advancements in MRI techniques include different aspects of brain structure such as white-matter tract integrity (measured with diffusion-tensor imaging) and voxel-based, volumetric assessments. In a cohort of elective surgery patients without dementia, white matter hyperintensities (WMHs) and whole-brain and hippocampal volume indices measured at baseline were not predictive of POD.^98^ Another study found no association between brain volumes and WMH with POD delirium incidence in a group of patients without dementia, although patients who developed delirium had a lower baseline MMSE (p = 0.01).^99^ Furthermore, gray matter volume decreases more in patients who experience POD compared with controls in the 3-month period following an episode of delirium, suggesting that episodes of delirium may precipitate neurodegeneration.^100^ In a study of patients with respiratory failure or shock admitted to an ICU, MRI obtained at hospital discharge and at 3 months of follow-up found that the duration of delirium correlated with cortical atrophy globally and in specific areas such as the superior frontal lobe and hippocampus. This volumeloss was associated with poorer cognitive performance at 12 months.^101^ While the above studies found that baseline structural measures did not predict delirium incidence, follow-up scans in patients who experienced POD offered preliminary data supporting a hypothesis that POD is associated with subsequent neurodegeneration.^117^
In contrast, other studies that included a subgroup with dementia found that white matter changes observed on baseline CT or MRI studies predicted the incidence of delirium during hospitalization in the 5-year period following imaging.^102^ Earlier work in smaller populations found that severe WMHs were associated with postoperative delirium (OR: 3.9; 95% CI: 1.2–12.5),^103^ and deep subcortical WMH was associated with delirium (OR: 3.04, 95% CI: 1.14–8.12; p = 0.027) and overall gray matter reduction,^104^ as well as atrophy in regions such as the temporal lobes and limbic areas.^105^ Others have shown an association between lateral ventricle size and the incidence of POD (OR: 3.23 per tertile increase in ventricular size; 95% CI: 1.21–8.60; p = 0.02).^106^ These studies either did not report on dementia diagnosis or excluded those with dementia or significant cognitive impairment, and were done in cardiac surgery patients; thus, they have limited generalizability.^103–106^
Given the evolving understanding of the relationship between dementia and delirium, the cortical thinning characteristic of an “AD signature” was examined in a cohort of patients undergoing elective surgery. While the “AD signature” did not predict delirium, it was associated with delirium severity.^107^ When these patients were investigated over a 36-month period following surgery, those with “AD signature” atrophy had a greater decline in general cognitive performance composite and those who also experienced delirium performed worse on a verbal learning test.^108^ A study of nondemented patients admitted to the ICU who developed delirium had accelerated gray matter thinning in temporal, frontal, and parietal regions compared with those who did not have delirium.^109^ Finally, using a combination of functional and structural imaging, a small study found that POD was independent of brain-amyloid pathology on PET imaging and these patients had a unique pattern of cortical atrophy.^110^ These studies highlight that delirium is associated with preexisting cortical atrophy and may accelerate the development of neurodegeneration. The relative vulnerability of particular patient populations (i.e., AD patients) remains undetermined.
Delirium is recognized as a disorder involving network disconnection.^1,111^ In patients with intracerebral hemorrhage (ICH), disruption of white matter tracts, for example, the right-hemispheric superior longitudinal fasciculus and those within the parahippocampal gyrus, is associated with increased delirium incidence.^112^ In acute ischemic stroke patients, age, previous stroke, and left-sided cortical strokes have been found to be independent predictors of POD.^113^ The thalamus, a critical network node in many corticostriatal loops, is increasingly implicated in POD,^114^ with increased mean diffusivity (MD) of the thalamus and particular thalamic nuclei shown to be associated with POD.^115^ Abnormalities of the thalamus, in addition to the cerebellum and hippocampus, are predictive of delirium incidence and severity.^116,117^ Finally, an absence of atrophy in the mid-cingulate cortex associated with a “SuperAging” phenotype appears to be protective against POD.^118^
Overall, the evidence relating structural imaging to the risk of POD continues to evolve. The predictive ability of any one structural abnormality is limited; however, combinations of structural information such as pathology-specific patterns or combinations of modalities may improve sensitivity. Furthermore, current data often lack sufficient patient data regarding dementia status and thus our ability to generalize the current data is limited. Future studies with larger patient populations should include those with and without dementia and ideally have MRI measures at multiple time points; so, we can better determine the interaction between delirium and subsequent risk for dementia.
Measures of brain perfusion (i.e., arterial spin-labeled [ASL], MRI, CT perfusion, and cerebral Doppler) and metabolic activity (FDG-PET, fMRI) have been used to examine the relationship between delirium and dementia. Decreased perfusion in the precuneus and posterior cingulate measured by ASL perfusion MRI was associated with poorer performance on a cognitive battery, but there was no predictive ability for delirium incidence.^119^ In an FDG-PET study, delirium was associated with hypometabolism globally, and increases in posterior cingulate cortex (PCC) FDG signal on delirium resolution were associated with better cognitive performance.^120^ The authors highlight that the PCC metabolism did not normalize; however, in the absence of pre-delirium imaging, no causal relation could be inferred. Other regions implicated in delirium pathogenesis include the thalamus, which the authors posit was a unique signature compared with known patterns in dementia.^121^ In a cohort of 44 patients admitted to a geriatric service with delirium and or dementia, cerebral blood flow velocity in the middle cerebral artery was measured using transcranial Doppler, and was able to distinguish between patients with delirium and patients with DSD. All patients with delirium had decreased flow velocities, but those with DSD had decreased CBF compared with the other groups.^122^
Resting-state functional MRI in nondemented patients during an episode of delirium revealed disturbances in network–network interactions.^123^ In particular, the normal anticorrelation observed between the default-mode network and the executive-attention network disappeared. This finding was later replicated in nondemented patients with additional nodes included in the analysis.^124^ Of note, these studies excluded patients with cognitive impairment, and did not include long-term imaging or cognitive testing follow-up, thus limiting inferences regarding the overlap between delirium and dementia.
Other work examining connectivity changes in delirium with fMRI found that delirium incidence and duration are associated with decreased neural network efficiency.^125^ Furthermore, the connectivity strength of the neural network was less efficient following an episode of delirium, supporting the notion of long-term network alterations following delirium.^125^ In a prospective, multicenter, observational cohort using similar fMRI analysis, decreased global connectivity following surgery persisted for 3 months in patients with POD, and decreased connectivity was associated with impaired performance on Trail Making Test B (TMT-B).^126^ Thus, disturbances in functional networks may partially account for residual cognitive deficits following delirium. Interestingly, predisposing risk factors for delirium such as vascular disease, depression, cognitive impairment, and physical status did not predispose to similar network dysfunctions as occur during delirium, suggesting that the network aberrancies occurring during delirium and potentially afterward may represent a unique state.^127^
In summary, functional neuroimaging has provided insight into the neurophysiological basis for delirium. Network integrity is crucial for normal brain function and interruptions of these networks have been associated with delirium. As with structural neuroimaging, future work should aim to include both patients with and without dementia to better understand the relation between these conditions.
As discussed earlier, there is growing evidence supporting the interrelationship between delirium and dementia, with delirium associated with accelerating cognitive decline and increasing the risk of subsequent dementia. Furthermore, delirium severity may impact cognitive decline in patients without dementia, as a prospective cohort of older adults undergoing major elective noncardiac surgery demonstrated a dose–response relationship, with the patients with the highest delirium severity experiencing the greatest rate of cognitive decline.^128^ Prevention of delirium, or even reduction in severity, might therefore lead to a decreased risk of dementia and the rapidity of cognitive decline. This is in line with recent attention to addressing modifiable risk factors to prevent dementia.^129^ Using a decision analysis approach, it was recently estimated that multicomponent nonpharmacologic delirium prevention strategies could prevent six new cases of ADRD per 1,000 patients over a 2.4-year period after hospital discharge.^130^
Delirium prevention includes both pharmacologic and non-pharmacologic strategies.^6^ Multicomponent interventions have been demonstrated to be effective in reducing the incidence of delirium.^131,132^ A recent systematic literature review of prevention strategies in older adults with dementia found seven studies,^132^ three of which were identified as moderate quality.^133–135^ Delirium-friendly preprinted postoperative orders (which addressed factors such as pain management, sleep augmentation, and prevention of UTIs, among other factors) compared with usual care resulted in a 37% absolute risk reduction (intervention: 60%, 97%, p < 0.001) in postoperative delirium in patients with preexisting dementia.^133^ In an RCT in patients with dementia and hip fracture, multidisciplinary interventions of staff education, active prevention, and rigorous monitoring for postoperative complications resulted in decreased rates of postoperative delirium (among other complications) as well as a higher percentage in the intervention group regaining previous ADLs at 12-month follow-up.^134^ A single 2-year, open-label study comparing rivastigmine with aspirin in patients with vascular dementia found that significantly fewer participants in the rivastigmine group developed delirium.^135^ Another trial of 62 patients with cognitive impairment randomly assigned patients to receive a rivastigmine patch for 3 days before to 7 days after surgery, and found that the intervention reduced both the incidence and the severity of POD.^136^ Other acetylcholinesterase inhibitors such as donepezil are associated with improved outcomes in critically ill patients with dementia,^137^ possibly with greatest benefit in those with cholinergic deficiency (i.e., dementia patients).^138^ Anticholinergic burden (ACB) among patients with DSD in a postacute care setting was found to be associated with poor attention, impaired working memory, decreased physical function, and longer length of stay.^139^ A Cochrane review of ACB in patients with MCI or dementia found low-certainty evidence that ACB was associated with increased risk of death, but no firm conclusions could be drawn for neuropsychiatric disturbances.^140^ PWD often use many prescription and over-the-counter medications, and alterations to those should be made with care.^141^
Occupational therapy has previously been shown to prevent delirium and improve functional outcomes in elderly, nonventilated ICU patients.^142^ At least two studies have examined the role of occupational therapy in DSD, but neither had control groups and thus served more as a demonstration of feasibility.^143,144^ Cognitively stimulating activities seem to be insufficient for DSD treatment,^145^ but may have some benefit for prevention. A recent review of available literature in 2023 concluded that currently available data are not sufficient to offer definitive recommendations regarding nonpharmacological prevention or the management of DSD.^146^
Table 6 highlights some recommendations for delirium prevention in PWD adapted from the Hospital Elder Life Program (HELP)^147^; however, the effectiveness of these interventions has not been fully studied. Moving forward, a concerted effort should be made to develop concise and easy-to-implement protocols for dementia patients across clinical settings. To this point, studies such as the PREvention Program for Alzheimer’s RElated Delirium (PREPARED), a cluster randomized trial investigating the effect of a multicomponent intervention to reduce the incidence, severity, duration, and frequency of DSD,^148^ and others will be crucial to develop evidence-based guidelines for the prevention of this devastating condition.
The identification of delirium can be challenging, even more so inpatients with dementia and especially those in the moderate and severe stages of dementia. The key to recognizing DSD is understanding the timeline or acuity of the change in mental status, and how the cognitive status is different from the individual’s baseline (Table 1). Nevertheless, if any uncertainty exists, a workup for underlying causes of delirium should be conducted. More widespread education about DSD will be crucial in improving awareness of this condition among healthcare providers. Delirium is a common syndrome in hospitals and encountered in many other settings, and given the serious consequences ofdelirium, strategies to prevent and treat delirium are greatly needed. Preclinical and clinical studies suggest an association of systemic and neuroinflammation with neuronal injury in delirium pathogenesis. Studies have shown the combination of increased brain vulnerability, an unmasking of undiagnosed or unrecognized dementia, and diminished brain reserve may be involved in the interrelationship between delirium and dementia. Despite the increasing amount of evidence on the overlap between delirium and dementia and studies implicating delirium prevention as a potential modifiable risk factor for dementia, there remain large gaps in our knowledge on how to best prevent and manage this condition. Additionally, very limited data are available on the broader impact of delirium beyond acute hospitalization and other nonacute settings. More research is needed to further our understanding of the overlap between delirium and dementia because of the many detrimental outcomes that follow with these conditions and the potential to impact cognitive decline. Further investigation into the neurobiology underlying delirium is also needed, as filling that gap in our knowledge will likely allow for targeted management and interventions. Understanding the relationship between delirium and dementia may lead to innovative management or treatments for both of these conditions which account for the most common neurocognitive disorders in older adults.
Mrs. A is an 81-year-old woman with a diagnosis of early AD. She otherwise is healthy with no other chronic conditions and was residing in the community by herself with some assistance from her children with a few instrumental activities of daily living, specifically finances, driving, and shopping. Despite her mild memory problems, she had been fairly independent. Her children started noticing mental status changes and took her to the ED. In the ED, she was ruled out for stroke but had persistent confusion, day/night reversal, and worsening short-term memory. After this first admission, she was discharged to a rehabilitation center and after a short stay returned home as she was thought to be at her baseline. However, she had a few similar episodes soon after. At the time of her third ED visit within a 6-month period, it was recognized that she was having repeated delirium episodes. During her last hospital admission, she presented with delirium on arrival to the ED and remained persistently delirious for several days. Her delirium episodes during the hospitalization included hyperactive delirium where she pulled her IV line out and received an antipsychotic medication. It was ultimately decided that it would not be possible for her to return home. She was discharged to a rehabilitation center. She was no longer able to ambulate independently and after her short rehabilitation stay was permanently placed in a nursing facility.
Her family wants to discuss why she had such a fast downward What are some of the things you can draw from her case?What are some of the available tools to identify delirium in a patient with a history of dementia?How would you have handled her care differently?What things would be on your differential diagnosis for triggering her delirium?What management for delirium prevention would you recommend to her inpatient and outpatient teams?