Journal of Psychosomatic Research 65 (2008) 239 – 248

Neuroimaging studies of delirium: A systematic review Roy L. Soiza a,b,⁎, Vijay Sharma b , Karen Ferguson c , Susan D. Shenkin d , David Gwyn Seymour a,b , Alasdair M.J. MacLullich d,e a

Department of Medicine and Therapeutics, University of Aberdeen, Aberdeen, Scotland, UK b Department of Medicine for the Elderly, Woodend Hospital, Aberdeen, Scotland, UK c SFC Brain Imaging Research Centre, Division of Clinical Neurosciences, University of Edinburgh, Western General Hospital, Edinburgh, Scotland, UK d Geriatric Medicine, University of Edinburgh, Royal Infirmary of Edinburgh, Edinburgh, Scotland, UK e MRC Centre for Cognitive Ageing and Cognitive Epidemiology, University of Edinburgh, Queen's Medical Research Institute, Edinburgh, Scotland, UK Received 20 February 2008; received in revised form 3 May 2008; accepted 15 May 2008

Abstract Objective: Neuroimaging offers clear potential in developing a better understanding of the pathophysiology of delirium. We performed a systematic review of structural and functional neuroimaging findings in delirium. The aims were to categorize and summarize the existing literature, and to determine whether this literature provides conclusive information on structural or functional brain predictors, correlates, or consequences of delirium. Methods: Studies were identified by comprehensive textword and MeSH-based electronic searches of MEDLINE, EMBASE, and Evidence-Based Medicine reviews, combining multiple terms for neuroimaging, brain structure, and delirium. Results: Twelve studies met the inclusion criteria. There were a total of 194 patients with delirium and 570 controls. Patient age, population, comorbidities, and identified precipitating factors were heterogeneous. Of the 10 structural studies, 3 studies used computed tomography (CT),

3 studies used magnetic resonance imaging (MRI), and 4 studies used a mixture of CT and MRI. One functional study used xenon CT, and the other used single photon emission computed tomography. There was a wide range of measurement techniques and timing of scans. Some studies found associations between delirium and cortical atrophy, and between ventricular enlargement and white matter lesion burden, but many studies did not control for potential confounders. Only two small studies of cerebral blood flow were identified, with both suggesting that there may be reduced regional cerebral blood flow, but the data were limited and somewhat inconsistent. Conclusions: The small sample sizes and other limitations of the studies identified in this review preclude drawing any clear conclusions regarding neuroimaging findings in delirium, but these studies suggest multiple avenues for future research. © 2008 Elsevier Inc. All rights reserved.

Keywords: Acute confusional state; Brain; Computed tomography; Delirium; Imaging; Magnetic resonance imaging

Introduction Delirium is a common and serious acute neuropsychiatric syndrome affecting between 10% and 41% of older medical patients [1]. In the last two decades, there has been Preliminary findings of this systematic review were presented at the regional scientific meeting (Scottish branch) of the British Geriatrics Society in Carnoustie on May 18, 2007, and at the national scientific meeting of the British Geriatrics Society in Glasgow on April 23–25, 2008. ⁎ Corresponding author. Department of Medicine and Therapeutics, University of Aberdeen, Polwarth Building, Foresterhill, Aberdeen AB25 2ZD, UK. Tel.: +44 1224 556319; fax: +44 1224 556339. E-mail address: [email protected] (R.L. Soiza). 0022-3999/08/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.jpsychores.2008.05.021

considerable progress in characterising the phenomenology, predisposing and precipitating factors, and outcomes of delirium [2–8]. It is now established that delirium is independently associated with multiple adverse outcomes, including higher morbidity, loss of independence, institutionalization, dementia, and death. However, little is known about the pathophysiology of delirium, making the development of appropriate therapeutic interventions more difficult. Neuroimaging techniques are fundamental to research into almost every disorder involving the central nervous system (CNS), particularly dementia, stroke, schizophrenia, and mood disorders, but have been scantily applied in the study of delirium. The current neuroimaging research literature on

240

R.L. Soiza et al. / Journal of Psychosomatic Research 65 (2008) 239–248

delirium spans many different patient populations, modalities, and methods of analysis. To help provide a basis for future work, we performed a systematic review. We excluded studies focusing on hepatic encephalopathy and delirium tremens because of the specific aetiologies and other features of these conditions. Butterworth [9] has previously summarized the relevant literature in hepatic encephalopathy where hyperintensities in the globus pallidus are particularly characteristic. It is thought that frontal and temporal atrophy predispose to delirium tremens [10]. The aims of the review were (a) to summarize the available literature and (b) to determine whether any conclusions regarding structural and functional neuroimaging predictors, correlates, or consequences of delirium could be drawn.

Methods Because all the studies were observational in design, we adhered to a protocol developed from a widely recommended method for systematic review/meta-analysis of observational studies (MOOSE) [11]. Study identification Two authors (R.L.S. and V.S.) sought studies published up to the end of 2007 that used any neuroimaging technique to investigate delirium. Studies were identified by comprehensive textword and MeSH-based electronic searches of MEDLINE, EMBASE, and Evidence-Based Medicine reviews (including the Cochrane database of systematic reviews and the Cochrane central register of controlled trials), combining terms for neuroimaging, brain structure, and delirium (for full search strategy, see Appendix A). Additionally, we hand searched and crossreferenced the bibliographies of relevant articles [11,15,16] and books. Selection We included all studies from which we could extract data on neuroimaging findings in patients with delirium, regardless of study design. We limited the review to articles written in the English language. We excluded single case reports and studies focusing on delirium tremens and hepatic encephalopathy because of the specific aetiologies and other features of these conditions. We also excluded articles that did not use one of the following diagnostic criteria for delirium or acute confusional state: current or previous versions of Diagnostic and Statistical Manual of Mental Disorders (Diagnostic and Statistical Manual of Mental Disorders, Third Edition; Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revised; or Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition), International Classification of Diseases, or validated measures based on these criteria such as the Confusion Assessment Method [14].

Data extraction and synthesis V.S. and R.L.S. extracted the data, which were checked by A.M.J.M. Data on study design, provenance, and type of imaging modality used were collected. We also recorded the aetiology of delirium where this was available. For cohort studies, we determined whether intraindividual comparisons of paired scans were made. As far as possible, we obtained numerical data, although outcome measures were largely categorical in case series (e.g., proportion of patients with lesions of the basal ganglia) or expressed as group differences in controlled studies. We also noted the probability (P values) of any observed differences having arisen by chance if the null hypothesis were true (i.e., that there was no difference between the scans of delirious and nondelirious subjects). Where this was not reported, the P value was calculated using Fisher's Exact Test for the difference between two proportions, and using unpaired t test for the difference in reported means. We assumed that variables were normally distributed where results were reported as means with standard deviations. Statistical significance was assumed when P≤.05. No quantitative data synthesis was performed due to the small number of studies identified and the heterogeneity of study populations and outcome measures. Assessment of study quality Two authors (R.L.S. and A.M.J.M.) independently assessed all manuscripts that met the selection criteria for quality. Details of the criteria used are shown in Table 1. There are no standard validated methods for the assessment of the quality of observational studies, especially those that involve different types of study design such as cohort and case–control studies. We therefore measured quality using a set of criteria developed for this review. These criteria were based as much as possible on published checklists produced to evaluate the design and reporting of similar types of studies [12,13]. Studies were given a point for each of the quality criteria present, and the sum of these generated an overall quality score out of 11.

Results On the basis of the initial search terms, 2191 citations were identified, with 57 articles retrieved for detailed evaluation after examination of the title and/or abstract. Articles not retrieved mostly concerned case reports, or studies on delirium tremens or hepatic encephalopathy. Twelve articles met the inclusion criteria for the systematic review (see Fig. 1). Table 1 provides details of the included studies. Excluded studies are listed in Appendix B. The main reasons for exclusions were that the studies: did not present original data (n=24), did not have an acceptable definition of delirium or acute confusional state (n=11), had data

R.L. Soiza et al. / Journal of Psychosomatic Research 65 (2008) 239–248 Table 1 Quality criteria Criterion

Description

A

Main aim of the study is to identify neuroimaging features of delirium Prospective design Sampling bias minimized; at least one third of potential volunteers recruited Delirium cases diagnosed by valid criteria For case–control studies: nondelirious control subjects with similar characteristics For cohort studies: appropriate information on subject characteristics, including possible confounders Use of validated methods of scans assessment Adequate quality control of scan results (e.g., use of two independent raters or reporting of interrater reliability) Blinding of scan reporter (i.e., no volunteer information available to the reporter) Scan report detail: localization and quantification of regional blood flow or the presence of at least two of the following: cerebral atrophy, white matter lesions, and pathological abnormalities Use of appropriate statistical methods to compare delirious with nondelirious subjects Inclusion of appropriate power calculation

B C D E

F G H I

J K

presented on single cases only (n=5), or did not report which imaging abnormalities were associated with delirium (n=5). Articles without original data were mostly editorial or review articles that did not present novel (unpublished) imaging findings. The relatively low number of case reports was due to earlier exclusion of those citations that had titles and/or abstracts that clearly indicated that the article was a case report. An overview of study designs, imaging techniques, and study settings is now provided, followed by additional details of included individual studies where relevant information is required for clarification. Overview The 12 articles meeting inclusion criteria comprised four cohort studies, five case–control studies, and three case series. Because our objective was to determine neuroimaging correlates of delirium and not to infer causality, we included all study types. One of the case series included functional imaging during and after delirium, such that the second scan acted as a “control” for the first [17]. One case–control study [18] also included a substudy of similar design with functional imaging during and after delirium. Four of the included studies did not set out to discover neuroanatomical correlates with delirium as the primary study objective. Nevertheless, these were retained because they met all inclusion criteria. Most studies employed computed tomography (CT) in all subjects or in an unspecified proportion of the subjects. One of these studies used xenon inhalation, allowing computation of regional blood flows. Two studies used magnetic resonance imaging (MRI) exclusively. One study employed

241

single photon emission computed tomography (SPECT). Therefore, the review includes 10 studies employing structural scans and 2 studies of regional cerebral blood flow. Studies from the United States (n=6), Europe (n=3), Australia (n=1), and Japan (n=2) were identified. Most studies were set in psychiatric departments (n=6). One of these studies only recruited patients with a previous stroke. Three further studies were set in stroke units and/or only recruited patients with stroke. Two studies were set in intensive care units. Only one study [18] recruited subjects from general medical wards, where most patients with delirium are usually assessed in routine clinical practice. Many studies only included delirium secondary to specific causes such as stroke and electroconvulsive therapy (ECT). Most studies included reasonably even numbers of male and female subjects, and mean ages ranged from 47.5 to 82.1 years. The mean ages were much lower in studies set in intensive care units. The median quality score was 5 out of 11 (see Table 2). A summary of the main descriptors and the results of these studies can be found in order of publication (earliest publication first) in Table 3. None of the studies included any power calculation, and many had small sample sizes. Studies ranged in size from 5 to 69 patients with delirium (and from 11 to 197 controls). A total of 194 cases and 570 controls are included in the review. Additional information on individual studies Koponen et al. [19] undertook CT brain scans on patients in psychogeriatrics wards who were experiencing delirium secondary to a range of causes and on a control group with neurological symptoms but without cognitive decline. The main outcome measures were (apparently) nonvalidated inhouse indices of the sizes of ventricular spaces relative to the diameter of the skull, and the nature and location of any regional changes. Patients with delirium had significantly higher indices of ventricular enlargement and an excess of low attenuation, particularly in the parieto-occipital lobes. Importantly, a high proportion (54 out of 69) of patients had dementia. However, in a subgroup analysis of delirious patients with milder cognitive impairment (N=16; mean Mini-Mental State Examination: 24.25) versus controls (N=31; mean Mini-Mental State Examination: 26.2), three out of the five measures of the brain remained significantly different. The authors concluded that structural brain disease predisposes to delirium. Figiel et al. [20] prospectively examined MRI predictors of delirium in 60 patients aged over 45 years who were commencing the use of antidepressants. The five incident cases of delirium were aged over 60 years and all had basal ganglia lesions, compared with just 9 out of 55 in those without delirium. Additionally, all cases had periventricular and deep white matter hyperintensities, and cortical atrophy or ventricular enlargement, compared with 21 out of 55 (38%) nondelirious subjects. Post-hoc analyses by the present

242

R.L. Soiza et al. / Journal of Psychosomatic Research 65 (2008) 239–248

Fig. 1. Flowchart of the selection of studies for inclusion in this review.

authors suggest that these differences were significant (Fisher's Exact Test, Pb.01). The same group then reported on a prospective cohort study in which consecutive older patients with depression who were undergoing ECT had MRI brain scanning prior to ECT [19]. They found higher numbers of basal ganglia lesions and white matter hyperintensities in those who developed delirium. Previously validated scales were used for lacunes and cortical atrophy [29] and for white matter lesions [30]. Figiel et al. [22] also reported similar findings in a case series (N=6) where MRI scans were

performed after the onset of ECT-induced delirium. The fourth report from this group [23] prospectively studied the incidence of ECT-induced delirium in 14 patients with depression after stroke, and neuroimaging findings (from CT and MRI) in cases and controls were compared. The four cases that developed delirium had caudate infarcts compared with 3 out of the 10 controls (P=.07). White matter lesions were not evaluated in this study. Four studies focused on stroke patients. In one case series, five patients with no premorbid cognitive impairment who

Table 2 Results of quality assessment Study

A

B

C

D

E

F

G

H

I

J

K

Total

Koponen et al. [19] Figiel et al. [20] Figiel et al. [21] Figiel et al. [22] Martin et al. [23] Nagaratnam and Nagaratnam [24] Kishi et al. [25] Henon et al. [26] Yokota et al. [17] Caeiro et al. [27] Samton et al. [28] Fong et al. [18]

1 1 1 1 0 1 0 0 1 0 0 1

1 1 1 1 1 0 1 1 0 1 0 1

1 1 1 0 1 0 1 1 0 1 0 0

1 1 1 1 1 1 1 1 0 1 1 1

0 0 1 0 0 0 0 1 0 1 0 1

0 0 1 0 0 0 0 1 0 0 1 1

0 0 1 0 0 0 0 0 0 0 1 1

1 1 1 0 0 0 0 0 0 0 1 1

1 1 1 0 0 0 0 1 1 0 1 1

1 0 1 0 0 0 0 1 1 1 0 1

0 0 0 0 0 0 0 0 0 0 0 0

7 6 10 3 3 2 3 7 3 5 5 9

R.L. Soiza et al. / Journal of Psychosomatic Research 65 (2008) 239–248

developed delirium had right hemispheric subcortical infarctions [24]. Henon et al. [26] examined the influence of preexisting cognitive decline on the occurrence of delirium in stroke patients. Cases (N=49) had greater degrees of cortical atrophy (P=.004) and white matter lesions (P=.045); both outcomes used standard scales. However, cases were older (median age: 78 vs. 74 years; P=.007). There was only one study involving stroke patients that stated one of its main aims to be the identification of neuroimaging correlates with delirium [27]. However, their analysis of images was limited to the nature and the location of the stroke. They found a higher proportion of cerebral hemisphere (as opposed to brainstem or cerebellum) strokes in patients with delirium (P=.02), but the small number of patients (n=29) and the wide variety of locations in which these strokes were found meant that no further conclusions could be reached. Furthermore, delirious patients were again significantly older, and age was a powerful uncorrected confounder. The study by Martin et al. [23] is described above. Kishi et al. [25] undertook the largest study (in terms of the total number of subjects) included in our review. This was an observational study of 325 consecutive patients admitted to a critical care unit, with the principal aim of elucidating the prevalence of delirium in this population. Most (n=235) underwent brain imaging, and 38 were diagnosed with delirium. Neuroimaging data were limited to abnormalities requiring medical or surgical intervention or observation. Cases had a rate of “pathological abnormalities” similar to that of controls. Samton et al. [28] reported CT findings in relation to Clock Drawing Test performance in 70 general hospital inpatients referred to a liaison psychiatry service. Twenty patients had a diagnosis of delirium. Neuroimaging data comprised the degree of subcortical atrophy estimated by calculation of the intercaudate ratio [31] and periventricular white matter lesion burden according to the method of van Swieten et al. [32]. No differences were observed between those with delirium and those with other diagnoses (depression, dementia, or “other”). However, demographic data were presented for the whole group, rather than subdivided by diagnosis; thus, confounding factors such as age could not be analysed. Two studies of cerebral perfusion were included. Using xenon CT, Yokota et al. [17] found that, during the episode of delirium, there were bilateral reductions in overall blood flow within all the studied cortical areas (frontal, temporal, and occipital) and deep grey matter, specifically the caudate head, thalamus, and lenticular nuclei. It should be noted that the subjects in this study were particularly young. Using SPECT, Fong et al. [18] found qualitative evidence of regional cerebral hypoperfusion in 50% of the patients with delirium compared with healthy age-matched and handedness-matched controls. They also reported significantly reduced regional blood flow ratios in the pons, left inferior frontal lobe, right temporal lobe, and right occipital lobe. In a

243

substudy of cases with paired scans (N=6), there were no visually detectable differences, but statistical parametric mapping revealed right parietal hypoperfusion in two subjects and left parietal hyperperfusion in another.

Discussion This is, to our knowledge, the first systematic review of studies of neuroimaging in delirium. We found 12 studies suitable for inclusion. Most had small sample sizes, and many used unsophisticated methods of imaging and analysis, or relied on clinical neuroimaging reports. The diversity of designs, patient populations, neuroimaging modalities, and types of measurement precludes any firm conclusions being drawn, particularly with respect to regional correlates of delirium. However, where these domains of measurement were performed, most studies of structural abnormalities found that patients with delirium had (a) more brain atrophy and (b) increased white matter lesions, although confounding by age and/or cognitive ability could not be excluded. Basal ganglia lesions were commonly reported as being associated with delirium, but in the absence of analysis of other brain regions, it is not certain whether this is a specific finding. The two functional studies suggest that delirium might be associated with perfusion abnormalities, but further work is required to determine whether these are global or whether specific regions are implicated. The findings of this review raise several issues. Perhaps the most striking is the paucity of studies: 12 were identified, and only 7 of these focused on delirium. The total number of cases in these seven studies was 127—a remarkably small number for a CNS disorder that affects at least 15% of general hospital inpatients [1,2]. Most structural studies lacked the rigor of typical neuroimaging studies of other CNS disorders. For example, no study gave an indication of the reliability of the methods used. In addition, quantitative measurements and validated white matter lesion scales were used infrequently. Even where validated scales were used, these varied among articles. This is important because published white matter lesion scales have different sensitivities; for example, the van Swieten scale was developed in patients with cerebrovascular disease and has a high ceiling effect when applied to patients with low to moderate lesion loads [33]. While advanced imaging techniques such as DTI, spectroscopy, functional MRI, and perfusion imaging have been used in stroke, dementia, and other psychiatric disorders, the methods used so far in investigating delirium have been basic and largely subjective. The generally insensitive modalities and/or methods of analysis mean that many types or more subtle degrees of abnormalities could have been missed. All but the functional studies were essentially concerned with predictors of delirium (i.e., abnormalities

244

Table 3 Main results

Design/setting

Koponen et al. [19], Finland

Case–control; psychiatric hospital (cases), neurological outpatient clinic (controls)

Figiel et al., [20] United States

Cohort; psychiatric hospital.

CT

MRI

Causes of delirium Stroke (22%) Metabolic (13%) Infection (13%) Medications (9%) Epileptic fit (9%) Life change in dementia (9%) Intracranial space-occupying lesion (9%) Carcinoma (6%) Myocardial infarction or insufficiency (4%) Functional psychosis (4%) Trauma (3%) Antidepressant

ECT

Mean age of cases (years)

Cases [n (male/female)]

73.5

69 (29/40)

72.4

76.1

5 (2/3)

10 (3/7)

Controls [n (male/female)] 31 (13/18)

55 (?)

Figiel et al. [21], United States

Cohort; psychiatric hospital.

MRI

77 (25/52)

Figiel et al. [22], United States Martin et al. [23], United States

Case series; psychiatric hospital Case–control; psychiatric hospital

CT/MRI

ECT

73.3

6 (3/3)

CT/MRI

ECT in stroke patients

68.4

4 (2/2)

Nagaratnam and Nagaratnam [24], Australia Kishi et al. [25], Japan

Case series; general hospital

CT

Stroke

72.2

5 (4/1)

0

Cohort; emergency admissions unit

CT/MRI

Trauma (29%) Cerebrovascular (18%)

54.9

38 (29/9)

197 (?)

0 10 (3/7)

Main positive findings indelirious patients

P

Higher frontal horn index Higher cella media index Greater width of third ventricle Greater width of sylvian fissure Greater mean of the largest cortical sulci Larger proportion of focal abnormalities (e.g., infarcts), especially on the right side Differences between hyperactive, hypoactive, and mixed delirium in:

b.001 b.001 b.001 b.01 .04 b.001

Frontal horn index Width of third ventricle

b.01 .03

All cases had basal ganglia lesions plus white matter hyperintensities and either cortical atrophy or ventricular enlargement Increased incidence of: Basal ganglia lesions Periventricular hyperintensity Deep white matter hyperintensity All cases had basal ganglia lesions and white matter hyperintensities Stroke involved caudate nucleus in 100% of delirious patients vs. 30% of nondelirious patients All cases had right hemisphere subcortical infarcts

b.001 a

No significant differences in pathological findings on scans

Quality score (/11) 7

6

.01 a 10 b.01 .02 .02 NA

3

.07 a

3

NA

2

.71 a

3

R.L. Soiza et al. / Journal of Psychosomatic Research 65 (2008) 239–248

Study/provenance

Imaging modality

Cohort; stroke unit

CT/MRI

Case series; intensive care unit

Xenon CT

Caeiro et al. [27], Portugal

Cohort; stroke unit

CT/MRI

Samton et al. [28], United States

Case–control; general hospital

CT

Fong et al. [18] United States

Case–control; general hospital

SPECT

Trauma (50%) Acute pancreatitis (30%) Perforated peptic ulcer (10%) Peritonitis carcinomatosa (10%) Stroke

Hypoxia/diffuse cerebral ischemia (25%) Systemic disease (50%) Drug intoxication or withdrawal (15%) Multifactorial: Medications (77%) Preexisting cognitive impairment (59%) Infection (50%) Dehydration (36%) Metabolic (36%) Hypoxia (14%) Immobility (11%)

78 (median) 47.5

49 (26/25) 10 (9/1) (all paired “before-and-after” scans)

63.0

29 (7/12)

Not stated

20 (?)

153 (71/82) 0

189 (113/76)

50 (?)

Greater cerebral atrophy score Greater white matter lesion burden Global reduction in cerebral blood flow Reduction in regional blood flow in all studied areas: right and left frontal, temporal and occipital lobes, and caudate head, thalamus, and lenticular nucleus

b.01 .05 b.01 All b.05

Higher proportion of hemispheric strokes relative to brainstem/ cerebellar strokes No significant differences in intercaudate ratio or periventricular white matter disease

.02

5

.99 b

5

22 (7/15) (6 paired “beforeand-after” scans)

11 (?)

3

.66 b

(total sample 32/38)

82.1

7

Decreased frontal, parietal, and occipital blood flow (qualitative data) Decreased regional blood flow ratios in the pons, left inferior frontal lobe, right temporal lobe, and right occipital lobe Paired scan results: Right parietal hypoperfusion (N=2) Left parietal hyperperfusion (N=1) Normal (N=3)

NA b.01

b.001 b.001 NA

9

R.L. Soiza et al. / Journal of Psychosomatic Research 65 (2008) 239–248

Henon et al. [26], France Yokota et al. [17], Japan

Acute abdomen (16%) Burn (8%) Circulatory failure (5%) Respiratory failure (3%) Poisoning (3%) Gastrointestinal bleed (3%) Stroke

(?) Data not present in published article. a b

P values not included in the original manuscript but derived from published data using Fisher's Exact Test. P values not included in the original manuscript but derived from published data using unpaired t test for delirium versus control subjects with depression.

245

246

R.L. Soiza et al. / Journal of Psychosomatic Research 65 (2008) 239–248

of the brain that increase the risk of delirium). Only the two functional studies involved serial imaging; both reported positive findings and, although sample sizes were small, offered useful models for further investigation of patients with current delirium. Another important issue was the lack of typical patients with delirium. Few of the included articles recruited patients from general medical or surgical populations, limiting the generalisability of their findings. Limitations This review was limited to studies published in the English language. We did not seek unpublished data. We excluded cases of delirium tremens and hepatic encephalopathy. It is always possible in any systematic review to inadvertently miss suitable articles. However, we screened 2191 citations, including several recent and older reviews, and we carefully hand searched bibliographies of these and other selected articles, so we believe that it is unlikely that any important publication has been omitted from this review. We short listed a relatively high number of articles only after hand searching bibliographies as we wished to be as thorough as possible, but all were later excluded. Metaanalysis of results was not possible due to the diversity of techniques and the small number of included studies. We performed no quantitative assessment of publication bias. We believe that the strict inclusion criterion of the definition of delirium is one of this review's strengths, but it is possible that some of the excluded articles contained genuine cases of delirium.

Conclusions The limitations in the studies identified in this review preclude the formation of any firm conclusions, but the findings of the studies are broadly consistent, there being associations between delirium and cortical atrophy, and between ventricular enlargement and increased white matter lesions. However, confounding from age, underlying disease, or other less obvious factors cannot be ruled out in the included studies. Associations between neuroimaging data and delirium do not necessarily imply that these changes are causal in the aetiology of delirium; they may simply indicate a more generally vulnerable brain. Future studies should employ standard diagnostic criteria and be powered to detect realistic effect sizes, which are probably small to moderate. Studies should also aim to test specific hypotheses, but it is likely that, at this early stage in the use of neuroimaging in delirium research, post-hoc findings will also suggest new lines of enquiry. Studies involving serial imaging before, during, and/or after episodes of delirium would be highly desirable, although we acknowledge the practical difficulties of imaging patients with delirium. Such studies will allow testing of the

possibility that delirium, at least with some aetiologies, reflects CNS damage. Delirium of different aetiologies should be compared, and whether subjects have hypoactive, hyperactive, or mixed delirium should be clearly reported. Control subjects should ideally be of similar ages and similar comorbidities. The importance of prior cognition and differences in brain structure (volume and white matter lesions) should be considered. The existing literature is small and has important methodological limitations, but has provided some preliminary information regarding vulnerability, process, and consequences of delirium. As interest in delirium research continues to grow, multimodal neuroimaging is likely to become a major tool in the investigation of the pathophysiology of this common yet poorly understood syndrome. Acknowledgments The UK Medical Research Council and the University of Edinburgh provided core funding for the MRC Centre for Cognitive Ageing and Cognitive Epidemiology, which supported this research. This work was supported by an MRC Clinician Scientist Fellowship to A.M.J.M. The authors wish to express their gratitude to the library staff at the NHS Grampian and at the University of Aberdeen Medical School for their help with seeking out and providing relevant manuscripts. Appendix A. Search strategy 1. Tomography, X-ray computed or CAT.mp or computed tomography.mp or computed axial tomography.mp or Magnetic Resonance Imaging/ or functional MRI.mp or Positron Emission Tomography or positron emission tomography.mp or PET.mp or Tomography, Emission-Computed, Single-Photon/ or single-photon emission tomography.mp or SPET.mp or SPECT.mp or CT.mp 2. White matter.mp OR Hippocampus OR hippocampus. mp or Leukoaraiosis/ or leukoaraiosis.mp OR prefrontal cortex.mp OR Prefrontal cortex/ OR intracranial.mp OR ventricle.mp OR morphometry.mp 3. Delirium/ OR delirium.mp OR acute confusional state. mp OR metabolic encephalopathy.mp 4. 1 or 2 5. 4 and 3 6. Limit 5 to English Appendix B. Excluded short-listed articles Alsop DC et al. The role of neuroimaging in elucidating delirium pathophysiology. J Gerontol Ser A Biol Sci Med Sci 2007;61:1287–93 Anonymous. Positron emission tomography (PET) and cortical localization: clinical studies in psychiatry and neurology. Psychopharmacol Bull 1982;18:2–8

R.L. Soiza et al. / Journal of Psychosomatic Research 65 (2008) 239–248

Bench CJ et al. Positron emission tomography in the study of brain metabolism in psychiatric and neuropsychiatric disorders. Br J Psychiatry 1990;9:82–95 Besson JA. Magnetic resonance imaging and its applications in neuropsychiatry. Br J Psychiatry 1990;9:25–37 Caplan LR et al. Caudate infarcts. Arch Neurol 1990;47:133–43 Coffey CE et al. Leukoencephalopathy in elderly depressed patients referred for ECT. Biol Psychiatry 1988;24:143–61 Devinsky O et al. Confusional states following posterior cerebral artery infarction. Arch Neurol 1988;45:160–63 Doyle M et al. Use of SPECT to evaluate postcardiotomy delirium. Am J Psychiatry 1996;153:838–39 Dunne JW et al. Inobvious stroke: a cause for delirium and dementia. Aust NZ J Med 1986;16:771–78 Fink M et al. ECT and delirium in Parkinson's disease. J Neuropsychiatry Clin Neurosci 1992;4:231–32 Friberg L. Regional cerebral blood flow in psychiatry as measured by single photon emission tomography. Psychiatry Res 1989;29:319–22 Gokgoz K et al. Psychiatric complications of cardiac surgery: postoperative delirium syndrome. Scand Cardiovasc J 1997;31:217–22 Graff-Radford NR et al. Non-haemorrhagic infarction of the thalamus: behavioral, anatomic and physiologic correlates. Neurology 1984;34:14–23 Gunaydin B et al. Cerebral hypoperfusion after cardiac surgery and anesthetic strategies: a comparative study with high dose fentanyl and barbiturate anesthesia. Ann Thorac Cardiovasc Surg 1998;4:12–17 Gustafson Y et al. Acute confusional states (delirium) in stroke patients. Cerebrovasc Dis 1991;1:257–64 Gustafson Y et al. Acute confusional state (delirium) soon after stroke is associated with hypercorticolism. Cerebrovasc Dis 1993;3:33–38 Horenstein S et al. Infarction of the fusiform and calcarine regions: agitated delirium and hemianopia. Trans Am Neurol Assoc 1967;92:85–89 Inouye SK et al. Elucidating the pathophysiology of delirium and the interrelationship between delirium and dementia. J Gerontol 2006;61A:1277–79 Irimia P et al. Delirium due to brain microembolism: diagnostic value of diffusion-weighted MRI. J Neuroimaging 2007;2:175–77 Koponen H et al. Acute confusional states in the elderly: a radiological evaluation. Acta Psychiatr Scand 1987;76: 726–31 Koponen HJ et al. Delirium among elderly persons admitted to a psychiatric hospital: clinical course during the acute stage and one-year follow-up. Acta Psychiatr Scand 1989;79:579–85 Koponen HJ et al. A prospective study of delirium in elderly patients admitted to a psychiatric hospital. Psychol Med 1993;23:103–9 Krausz Y et al. Brain SPECT imaging of neuropsychiatric disorders. Eur J Radiol 1996;21:183–87

247

Kumral E et al. Spectrum of anterior cerebral artery territory infarction: clinical and MRI findings. Eur J Neurol 2002;9:615–24 Lerner DM et al. Neuroimaging in delirium and related conditions. Semin Clin Neuropsychiatry 2000;5:98–112 Lipowski ZJ. Delirium updated. Compr Psychiatry 1980;21:190–96 McManus J et al. Delirium post-stroke. Age Ageing 2007;36:613–18 Medina JL et al. Agitated delirium caused by infarctions of the hippocampal formation and fusiform and lingual gyri. Neurology 1974;24:1181–83 Medina JL et al. Syndrome of agitated delirium and visual impairment: a manifestation of medial temporooccipital infarction. J Neurol Neurosurg Psychiatry 1977; 40:861–64 Mesulam MM et al. Acute confusional states with right middle cerebral artery infarction. J Neurol Neurosurg Psychiatry 1979;39:84–89 Moles JK et al. Neurological deficits and CT findings in psychiatric patients. Psychosomatics 1998;39:394–95 Mori E et al. Acute confusional state and acute agitated delirium. Arch Neurol 1987;44:139–43 Mullaly W et al. Frequency of acute confusional states with right middle cerebral artery infarction. Ann Neurol 1982;12:113 Naughton BJ et al. Computed tomography screening and delirium in elder patients. Acad Emerg Med 1997;12: 1107–10 Nicolai A et al. Acute confusional states secondary to infarctions in the territory of the posterior cerebral artery in elderly patients. Ital J Neurol Sci 1994;15: 91–96 Santamaria J et al. Confusional syndrome in thalamic stroke. Neurology 1984;34:1618 Schmidley JW et al. Agitated confusional states in patients with right hemisphere infarctions. Stroke 1984;15: 883–85 Schroeder J et al. Regional cerebral blood flow in endogenous psychoses: a Tc-99m HMPAO-SPECT pilot study. Psychiatr Res 1989;29:331–33 Sheng AZ et al. Delirium within three days of stroke in a cohort of elderly patients. J Am Geriatr Soc 2006; 54:1192–98 Shih HT et al. Confusion or delirium in patients with posterior cerebral arterial infarction. Acta Neurol Taiwan 2007;16:136–42 Trzepacz PT et al. The relationship of SPECT scans to behavioural dysfunction in neuropsychiatric patients. Psychosomatics 1992;33:62–71 Trzepacz PT. Update on the neuropathogenesis of delirium. Dementia Geriatr Cogn Disord 1999;10: 330–34 Trzepacz PT. Is there a final common neural pathway in delirium? Focus on acetylcholine and dopamine. Semin Clin Neuropsychiatry 2000;5:132–48

248

R.L. Soiza et al. / Journal of Psychosomatic Research 65 (2008) 239–248

Vasile RG. Single photon emission computed tomography in psychiatry: current perspectives. Harvard Rev Psychiatry 1996;2:27–38 Weinberger DR. Neuroimaging in neuropsychiatry. New Dir Ment Health Serv 1992;54:77–82

[17]

[18]

References [19] [1] Siddiqi N, House AO, Holmes JD. Occurrence and outcome of delirium in medical in-patients: a systematic literature review. Age Ageing 2006;35:350–64. [2] Young J, Inouye SK. Delirium in older people. Br Med J 2007;334:842–846D. [3] Marcantonio ER, Simon SE, Bergmann MA, Jones RN, Murphy KM, Morris JN. Delirium symptoms in post-acute care: prevalent, persistent, and associated with poor functional recovery. J Am Geriatr Soc 2003;51:4–9. [4] Meagher DJ, Moran M, Raju B, Gibbons D, Donnelly S, Saunders J, Trzepacz PT. Phenomenology of delirium—assessment of 100 adult cases using standardised measures. Br J Psychiatry 2007;190:135–41. [5] Ely EW, Shintani A, Truman B, Speroff T, Gordon SM, Harrell FE, Inouye SK, Bernard GR, Dittus RS. Delirium as a predictor of mortality in mechanically ventilated patients in the intensive care unit. JAMA 2004;291:1753–62. [6] McCusker J, Kakuma R, Abrahamowicz M. Predictors of functional decline in hospitalized elderly patients: a systematic review. J Gerontol Ser A—Biol Sci Med Sci 2002;57:M569–77. [7] Rockwood K, Cosway S, Carver D, Jarrett P, Stadnyk K, Fisk J. The risk of dementia and death after delirium. Age Ageing 1999;28:551–6. [8] Jackson JC, Gordon SM, Hart RP, Hopkins RO, Ely EW. The association between delirium and cognitive decline: a review of the empirical literature. Neuropsychol Rev 2004;14:87–98. [9] Butterworth RF. Pathogenesis of hepatic encephalopathy: new insights from neuroimaging and molecular studies. J Hepatol 2003;39:278–85. [10] Maes M, Vandoolaeghe E, Degroote J, Altamura C, Roels C, Hermans P. Linear CT-scan measurements in alcohol-dependent patients with and without delirium tremens. Alcohol 2000;20:117–23. [11] Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, Moher D, Becker BJ, Sipe TA, Thacker SB. Meta-analysis of observational studies in epidemiology—a proposal for reporting. JAMA 2000;283:2008–12. [12] Alsop DC, Fearing MA, Johnson K, Sperling R, Fong TG, Inouye SK. The role of neuroimaging in elucidating delirium pathophysiology. J Gerontol Ser A—Biol Sci Med Sci 2006;61:1287–93. [13] Lerner DM, Rosenstein DL. Neuroimaging in delirium and related conditions. Sem Clin Neuropsychiatry 2000;5:98–112. [14] Inouye SK, Vandyck CH, Alessi CA, Balkin S, Siegal AP, Horwitz RI. Clarifying confusion—the Confusion Assessment Method—a new method for detection of delirium. Ann Intern Med 1990;113:941–8. [15] Rothwell PM, Pendlebury ST, Wardlaw J, Warlow CP. Critical appraisal of the design and reporting of studies of imaging and measurement of carotid stenosis. Stroke 2000;31:1444–50. [16] von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PF, Vandenbroucke JP. The Strengthening the Reporting of Observational

[20]

[21]

[22]

[23]

[24] [25]

[26]

[27] [28]

[29]

[30]

[31]

[32]

[33]

Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Ann Intern Med 2007;147:573–7. Yokota H, Ogawa S, Kurokawa A, Yamamoto Y. Regional cerebral blood flow in delirium patients. Psychiatry Clin Neurosci 2003;57: 337–9. Fong TG, Bogardus ST, Daftary A, Auerbach E, Blumenfeld H, Modur S, Leo-Summers L, Seibyl J, Inouye SK. Cerebral perfusion changes in older delirious patients using 99mTc HMPAO SPECT. J Gerontol Ser A —Biol Sci Med Sci 2006;61:1294–9. Koponen H, Hurri L, Stenback U, Mattila E, Soininen H, Riekkinen PJ. Computed-tomography findings in delirium. J Nerv Ment Dis 1989;177:226–31. Figiel GS, Krishnan KRR, Breitner JC, Nemeroff CB. Radiologic correlates of antidepressant-induced delirium: the possible significance of basal ganglia lesions. J Neuropsychiatry Clin Neurosci 1989;1:188–90. Figiel GS, Coffey CE, Djang WT, Hoffman G, Doraiswamy PM. Brain magnetic resonance imaging findings in ECT-induced delirium. J Neuropsychiatry Clin Neurosci 1990;2:53–8. Figiel GS, Krishnan KRR, Doraiswamy PM. Subcortical structural changes in ECT-induced delirium. J Geriatr Psychiatry Neurol 1990;3:172–6. Martin M, Figiel GS, Mattingly G, Zorumski CF, Jarvis MR. ECTinduced interictal delirium in patients with a history of a CVA. J Geriatr Psychiatry Neurol 1992;5:149–55. Nagaratnam N, Nagaratnam P. Subcortical origins of acute confusional states. Eur J Intern Med 1995;6:55–8. Kishi Y, Iwasaki Y, Takezawa K, Kurosawa H, Endo S. Delirium in critical care unit patients admitted through an emergency room. Gen Hosp Psychiatry 1995;17:371–9. Henon H, Lebert F, Durieu I, Godefroy O, Lucas C, Pasquier F, Leys D. Confusional state in stroke—relation to preexisting dementia, patient characteristics, and outcome. Stroke 1999;30:773–9. Caeiro L, Ferro JM, Albuquerque R, Figueira ML. Delirium in the first days of acute stroke. J Neurol 2004;251:171–8. Samton JB, Ferrando SJ, Sanelli P, Karimi S, Raiteri V, Barnhill JW. The Clock Drawing Test: diagnostic, functional, and neuroimaging correlates in older medically ill adults. J Neuropsychiatry Clin Neurosci 2005;17:533–40. Coffey CE, Figiel GS, Djang WT, Cress M, Saunders WB, Weiner RD. Leukoencephalopathy in elderly depressed-patients referred for ECT. Biol Psychiatry 1988;24:143–61. Fazekas F, Chawluk JB, Alavi A, Hurtig HI, Zimmerman RA. MR signal abnormalities at 1.5-T in Alzheimer dementia and normal aging. AJR Am J Roentgenol 1987;149:351–6. Reider-Groswasser I, Ommaya AK, Salazar AM. Neuroimaging assessment in TBI and stroke: relevance for acute and post acute treatment. In: Leon-Carrion J, editor. Neuropsychological rehabilitation: fundamental, directions and innovations. Delray Beach: GR/St. Lucie Press, 1997. pp. 73–108. van Swieten JC, Hijdra A, Koudstaal PJ, van Gijn J. Grading white matter lesions on CT and MRI—a simple scale. J Neurol Neurosurg Psychiatry 1990;53:1080–3. Wardlaw JM, Ferguson KJ, Graham C. White matter hyperintensities and rating scales—observer reliability varies with lesion load. J Neurol 2004;251:584–90.