Delirium pathophysiology

Delirium Microchapters
Home
Patient Information
Overview
Historical Perspective
Classification
Pathophysiology
Causes
Differentiating Delirium from other Diseases
Epidemiology and Demographics
Risk Factors
Natural History, Complications and Prognosis
Diagnosis
Diagnostic Criteria
History and Symptoms
Physical Examination
Laboratory Findings
CT
MRI
Other Diagnostic Studies
Treatment
Medical Therapy
Primary Prevention
Secondary Prevention
Cost-Effectiveness of Therapy
Future or Investigational Therapies
Case #1
Delirium On the Web
Most recent articles
Most cited articles
Review articles
CME Programs
Powerpoint slides
Images
American Roentgen Ray Society Images of Delirium All Images X-rays Ultrasound CT Images MRI
Ongoing Trials at Clinical Trials.gov
US National Guidelines Clearinghouse
NICE Guidance
FDA on Delirium
CDC on Delirium
Delirium in the news
Blogs on Delirium
Directions to Hospitals Treating Delirium
Risk calculators and risk factors for Delirium

Editor-In-Chief: C. Michael Gibson, M.S., M.D. [1]; Associate Editor(s)-in-Chief: Vishal Khurana, M.B.B.S., M.D. [2] ; Pratik Bahekar, MBBS [3]

Pathophysiology

Cortical and subcortical dysfunction are behind development of the delirium. Reduced acetylcholine and histamine activity, increased dopamine and glutamate activity is observed in delirium. Role of GABA and serotonin is uncertain.

Animal models

The pathophysiology of delirium is not well understood and a lack of animal models that are relevant to the syndrome has left many key questions in delirium pathophysiology unanswered. Earliest rodent models of delirium used an antagonist of the muscarinic acetylcholine receptors, atropine, to induce cognitive and EEG changes similar to delirium. Similar anticholinergic drugs such as biperiden and scopolamine have also produced delirium-like effects. These models, along with clinical studies of drugs with ‘anticholinergic activity’ have contributed to a hypocholinergic theory of delirium.

Profound systemic inflammation occurring during bacteraemia/sepsis is also known to cause delirium (often termed septic encephalopathy). Modeling this in mice also causes robust brain dysfunction and probably a delirium-like state, although these animals are typically too sick to assess cognitively and measures such as EEG and magnetic resonance imaging/spectroscopy are necessary to demonstrate dysfunction.

Animal models that interrogate interactions between prior degenerative pathology and superimposed systemic inflammation have been developed more recently and these demonstrate that even mild systemic inflammation, a frequent trigger for clinical delirium, induces acute and transient attentional/working memory deficits, but only in animals with prior pathology. Prior dementia or age-associated cognitive impairment is the primary predisposing factor for clinical delirium and ‘prior pathology’ as defined by these new animal models may consist of synaptic loss, network disconnectivity, and primed microglia (brain macrophages that are ‘primed’ by the primary pathology to produce exaggerated responses to subsequent inflammatory insults).

While it is difficult to state with confidence whether delirium is occurring in a non-verbal animal, comparisons with human DSM-IV criteria remain useful. According to DSM-IV, demonstration of acute onset impairments in attention and some other cognitive domain, that cannot be better explained by existing dementia and that are triggered by physiological disturbances resulting from some general medical condition should be present in order to reach a ‘diagnosis’ of delirium. Recent animal models fulfill these criteria reasonably well. Whether the deficit is one of attention or short-term memory is difficult to dissect, but it is undeniably distinct from long-term memory, consistent with observations in patients with delirium. There is an urgent need to understand more about the mechanisms of dysfunction underpinning delirium and data arising from these and other animal models must form part of the discussion on delirium pathophysiology.

Clinical studies

Cerebrospinal fluid biomarkers

Studies of cerebrospinal fluid (CSF) in delirium are difficult to perform. Apart from the general difficulty of recruiting participants who are often unable to give consent, the inherently invasive nature of CSF sampling makes such research particularly challenging. However, a few studies have exploited the opportunity to sample CSF from persons undergoing spinal anaesthesia for elective or emergency surgery. Indeed, spinal anaesthesia may in fact be the anaesthetic modality of choice for frail older patients, so these studies are often undertaken in highly relevant populations.

A systematic review identified 8 studies involving 235 patients (142 with delirium). Overall, 17 different biomarkers were considered and each article identified in the review focused on a narrow range of biomarkers with no overlap between studies. Studies were generally small, studying heterogeneous populations with different times of CSF sampling in relation to delirium, and no clear conclusions could be drawn. Broadly, delirium may be associated with: increased serotoninergic and dopamine signalling; reversible fall in somatostatin; increased cortisol; and increase in some inflammatory cytokines (IL-8), but not others (TNF-α, IL-1β).

One additional study has since been published. Postoperative delirium was strongly associated with pre-operative cognitive decline. However, CSF Aβ1-42, tau, and phosphorylated-tau levels were not associated with delirium status, nor did they correlate significantly with cognitive function before the onset of delirium. The two main explanations for these findings are either: (1) the study was underpowered to detect mediating pathways between premorbid cognitive impairment, Alzheimer’s pathology biomarkers and subsequent delirium; or (2) postoperative delirium arises through pathophysiological pathways that are distinct from Alzheimer's disease.

Neuroimaging

The neuroimaging correlates of delirium are very difficult to establish. Many attempts to image people with concurrent delirium will be unsuccessful. In addition, there is a more general bias selecting younger and fitter participants amenable to scanning, especially if using intensive protocols such as MRI.

Most of the literature has been summarised by a systematic review. This found 12 articles for inclusion, most with small sample sizes (total number of cases 127). There was substantial heterogeneity in populations, study design, and imaging modalities such that no firm conclusions were made. Generally, structural imaging suggested that diffuse brain abnormalities such as atrophy and leukoaraiosis might be associated with delirium, though few studies could account for differences in key variables such as age, sex, education or underlying cognitive function and education.

Since publication of the systematic review, five further studies have been published. The largest-scale report was VISIONS. This prospectively examined the neuroimaging correlates of delirium in 47 participants after critical illness. Delirium duration was related to measures of white matter tract integrety and this in turn was related to poorer cognitive outcomes at 3 and 12 months. In addition, brain volumes were also assessed and related to cognitive outcomes in the same manner. Overall, the study found that longer duration of delirium was associated with smaller brain volume and more white matter disruption, and both these correlated with worse cognitive scores 12 months later.

Two studies examined delirium risk as a post-operative complication after elective cardiac surgery. These both showed that white matter damage predicted post-operative delirium. One functional MRI study reported a reversible reduction in activity in brain areas localising with cognition and attention function.

Neurophysiology

Electroencephalography (EEG) is an attractive mode of study in delirium as it is able to capture measures of global brain function. There are also opportunities to summarise temporal fluctuations as continuous recordings, compressed into power spectra (quantitative EEG, qEEG). Since the work of Engel and Romano in the 1950s, delirium has been known to be associated with a generalised slowing of background activity.

A systematic review identified 14 studies for inclusion, representing a range of different populations: 6 in older populations, 3 in ICU, sample sizes between 10 and 50). For most studies, the outcome of interest was the relative power measures, in order: alpha, theta, delta frequencies. The relative power of the theta frequency was consistently different between delirium and non-delirium patients. Similar findings were reported for alpha frequencies. In two studies, the relative power of all these bands was different within patients before and after delirium.

Neuropathology

Only a handful of studies exist where there has been an attempt to correlate delirium with pathological findings at autopsy. A case series has been reported on 7 patients who died during ICU admission. Each case was admitted with a range of primary pathologies, but all had acute respiratory distress syndrome and/or septic shock contributing to the delirium. 6/7 had evidence of hypoperfusion and diffuse vascular injury, with consistent involvement of the hippocampus in 5/7.

A case-control study examined 9 delirium cases with 6 age-matched controls, investigating inflammatory cytokines and their role in delirium. Persons with delirum had higher scores for HLA-DR and CD68 (markers of microglial activation), IL-6 (cytokines pro-inflammatory and anti-inflammatory activities) and GFAP (marker of astrocyte activity). These results might suggest a neuroinflammatory substrate to delirium, but the conclusions are limited by biases from selection of controls.