- Email: [email protected]
Transcranial Doppler to Measure Cerebral Blood Flow in Delirium Superimposed on Dementia. A Cohort Study
JAMDA 15 (2014) 355e360
JAMDA journal homepage: www.jamda.com
Original Study
Transcranial Doppler to Measure Cerebral Blood Flow in Delirium Superimposed on Dementia. A Cohort Study Gideon A. Caplan MD a, b, *, ZhongZheng Lan BSc(Med) c, Lyndal Newton MBBS a, b, Tasha Kvelde DPsy a, Catherine McVeigh MD a, b, Mark A. Hill PhD c a
Department of Geriatric Medicine, Prince of Wales Hospital, Sydney, Australia Prince of Wales Clinical School, University of New South Wales, Sydney, Australia c School of Medical Sciences, University of New South Wales, Sydney, Australia b
a b s t r a c t Keywords: Delirium Alzheimer’s disease ultrasound assessment of cognitive disorders/dementia geriatric assessment
Objective: Delirium superimposed on dementia (DSD) is frequently not diagnosed, at great cost. Both delirium and dementia are associated with cerebral hypoperfusion. A switch to anaerobic glycolysis in the central nervous system during delirium compared to Alzheimer’s dementia (AD) suggests greater hypoperfusion in DSD. The main aims of this study were to investigate whether cerebral hypoperfusion could differentiate DSD from related entities, and the characteristics of that hypoperfusion. Methods: Prospective cohort study of 44 Geriatric Medicine patients in 4 groups: (1) delirium, no history of dementia; (2) DSD; (3) acute illness without delirium or dementia; and (4) AD, no delirium. We measured CBF using transcranial Doppler to assess flow velocity (FV) and pulsatility index in the middle cerebral artery (MCA). Results: DSD has lower FV than either AD or delirium alone, or acute illness (28.2 4.7 vs AD: 41.3 15.7; P ¼ .040; vs delirium 37.7 8.2; P ¼.009; vs acute illness 43.0 8.3; P <.001). A mean MCA FV cut-off of 32.25 cm/s diagnoses DSD with a sensitivity of 0.875 and specificity of 0.788, area under the curve 0.884; P ¼ .001. Resolution of delirium improves FV (P ¼ .005). FV correlates with delirium severity (delirium index R ¼ 0.39; P ¼ .009) and dementia (Mini-Mental State Examination, R ¼ 0.33; P ¼ .029, and Informant Questionnaire on Cognitive Decline in the Elderly, R ¼ 0.41; P ¼ .005). Conclusions: Transcranial Doppler is a potential diagnostic and monitoring test for DSD. Correlation with clinical indicators of delirium suggests pathophysiological significance. Crown Copyright Ó 2014 Published by Elsevier Inc. on behalf of American Medical Directors Association, Inc. All rights reserved.
Delirium is a serious complication of illness, commonly affecting older patients, including more than 20% of hospitalized people over 65, and 70%e87% of those in intensive care, as well as being a major problem in long-term care.1,2 It is a significant cause of preventable morbidity and mortality in older patients, substantially increasing healthcare utilization and costs.1,3 Diagnosis is clinical, importantly informed by the corroborative history, however, 85% of delirious patients remain undiagnosed in hospital and 50% in nursing homes, or are misdiagnosed as having dementia, despite delirium being a medical emergency.1,4 This is partly because of a lack of diagnostic This study was supported by an unrestricted grant from the Julia Lowy Foundation. Z.Z.L. was partially funded by a scholarship from the NSW Branch of the Australian and New Zealand Society for Geriatric Medicine. The authors declare no conflicts of interest. * Address correspondence to Gideon A. Caplan, MD, Department of Geriatric Medicine, Prince of Wales Hospital, Edmund Blackett Building, Randwick, New South Wales 2031, Australia. E-mail address: [email protected] (G.A. Caplan).
tests, especially when superimposed on dementia.5,6 Diagnostic indices are available, such as the Confusion Assessment Method (CAM), but it is unreliable in untrained hands.7 Delirium is more common in people with dementia, is the most common acute driver of dementia, with which it shares many risk factors, and causes worsening of cognition and performance of activities of daily living, the clinical measures of dementia.8,9 Delirium superimposed on dementia (DSD) is more frequently missed and results in higher mortality and readmission rates.10e12 We recently showed that cerebrospinal fluid lactate is elevated and neurone specific enolase decreased in delirium suggesting a switch to anaerobic glycolysis,13 compared with Alzheimer’s dementia (AD) where altered glucose metabolism is the most sensitive and specific marker of AD.14 Abnormal glycolysis could lead to widespread neuronal dysfunction, triggering the clinical manifestations of delirium. Changes in glycolysis could be due to excessive demand for energy, toxins or reduced cerebral blood flow (CBF). The brain is dependent on glucose for energy production and inadequate
1525-8610/$ - see front matter Crown Copyright Ó 2014 Published by Elsevier Inc. on behalf of American Medical Directors Association, Inc. All rights reserved. http://dx.doi.org/10.1016/j.jamda.2013.12.079
356
G.A. Caplan et al. / JAMDA 15 (2014) 355e360
CBF could disrupt glucose transport or its complete oxidation, thus lowering the efficiency of glucose metabolism.13 Studies of CBF in delirium have been inconclusive, with variable findings ranging from no change to hypoperfusion and hyperperfusion.15e17 Cerebral hypoperfusion predisposes to postoperative cognitive decline, which is linked to delirium.18 Many patients with delirium also have dementia so the issue is complicated by evidence that decreased CBF is characteristic of AD, which also predisposes to delirium. However, CBF in delirium and AD has never been compared, therefore, it is unclear if the changes seen in delirium are due to underlying AD.19 In addition, CBF declines by between 28% and 50% from age 30 to age 70.20 We investigated the hypothesis that DSD is associated with greater reductions in CBF by noninvasive, safe transcranial Doppler (TCD). Methods This cohort study was approved by the Human Research Ethic Committee. Written informed consent was obtained from all patients or their person responsible, where the patient lacked capacity. In 2011 we recruited consecutive consenting patients in four groups with (1) acute illness and delirium but no history of dementia; (2) acute illness with delirium and a history of dementia, the DSD group; (3) acute illness without delirium or dementia; and (4) no acute illness or delirium, known (Alzheimer’s dementia) AD. Groups 1e3 were recruited from the Acute Geriatrics inpatient ward and were diagnosed by geriatricians based on clinical experience and the CAM.21 Patients admitted to the Geriatrics ward are screened with the CAM in the Emergency Department. Patients with AD were recruited from both a general geriatric outpatient clinic and a cognitive disorders clinic, with AD diagnosis based on Diagnostic and Statistical Manual, 4th edition, and National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association criteria for probable Alzheimer’s dementia. Cognitive and psychological assessments were conducted by a geriatrician and a neuropsychologist. Patients were excluded if they had (1) possible impaired cerebral vasoreactivity because of (a) sonographic evidence of severe extracranial stenosis (>70%); (b) sonographic evidence of intracranial stenosis, and (c) presence of territorial infarcts larger than one third of the hemisphere; and (2) no insonation window or a poor signal. Patients with underlying dementia or history of delirium during current hospitalization were excluded from the acute illness group. Patients were first screened for the presence of an insonation window. After a TCD measurement was taken, all subjects underwent baseline assessments through the completion of the clinical measures listed below. No intervention was performed on any of the patients for this observational study. Assessments Cognitive and psychological assessments were conducted using the Mini-Mental State Examination (MMSE);22 the Confusion Assessment Method (CAM);21 the delirium index (DI);23 the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE);24 and the Geriatric Depression Scale (GDS).25 Physiological and functional status assessment were carried out using the Barthel index;26 the modified Instrumental Activities of Daily Living (IADL) index;27 the Acute Physiology, Age, Chronic Health Evaluation (APACHE) II;28 and the Charlson comorbidity index (CCI).29 The IQCODE was asked in regard to the patient’s status relative to the pre-hospitalization status. For patients in groups 1e3 who were in hospital, the Doppler and CAM were repeated second daily until discharge. No testing was done prior to hospitalization. Group 4 were assessed only once. The order of testing was the same for all groups. The results in Table 2 refer to the initial testing.
TCD Imaging We used the QL modular software and Doppler-Box from DWL Compumedics for analyzing transcranial spectral signals derived from systolic velocity, diastolic velocity, FV, and pulsatility indices generated by the MCA. The transtemporal approach of insonation was performed with the patient either lying supine on the examination bed or sitting upright in a chair and the head straight in both positions. The 2 MHz pulsed TCD probe was positioned according to the criteria of Aaslid,30,31 which allowed identification of the MCA. All measurements were stored on hard disk for off-line analysis. The average of right and left FV and PI were used for analyses if both MCA could be insonated adequately. A single measurement was made on every second day of their hospitalization for patients in the delirium and acute illness groups. A one-off reading was taken from all patients with AD from the outpatient clinic. All TCD readings were taken by one person (Z.Z.L.) to reduce variability. Statistical Analysis SPSS 21 for Windows (SPSS Inc, Chicago, IL) was used in this study. A 2-tailed P value of less than .05 was considered significant. For parametric data, t tests were used for univariate analysis, while Mann-Whitney U was used for nonparametric data. Multivariate analysis was performed with logistic and binomial regression. Fisher exact test was used to compare proportions. Receiver operating characteristics curve was calculated using a nonparametric distribution assumption. Missing data were excluded listwise. A sample size calculation indicated that this sample size would be sufficient to detect a 20% difference in CBF. Groups were compared on their initial TCD and CAM only. However, initial and postdelirium TCD was compared in groups 1þ2 with initial and final TCD in group 3. Results Between May and September 2011, a total of 86 patients were screened and 44 patients were eligible for inclusion and consented to participate in this study (Figure 1, online only). Patient characteristics (Table 1) were similar in the 3 groups except for age. Age was significantly lower in the AD group compared with the delirium [difference 9.5 years; 95% confidence interval (CI) 1.4e17.7; P ¼ .025], and acute illness groups (10.0 years; 95% CI 2.1e18.0; P ¼ .026). There were no significant differences in the reasons for hospitalization between the acute illness and delirium groups. There was a trend toward more patients with diabetes and more treated with psychotropics in the delirium group. There were significant differences among the 3 groups in the following measures: APACHE II, MMSE, CAM, DI, Barthel index, IQCODE, and modified IADL according to the expected clinical characteristics of each group (Table 2). However, the raised IQCODE in the delirium without dementia group may indicate that respondents had difficulty separating the cognitive decline associated with delirium from the premorbid state. The delirium and acute illness groups had similar acuity of illness on the APACHE II. The dementia and delirium groups had similar cognitive decline on the IQCODE, although the MMSE was lower in the delirium groups than in AD and acute illness, cognitive decline being a feature of both. The delirium groups scored more highly on both measures of delirium, the CAM and DI, than other groups. On univariate analysis mean, systolic and diastolic FV was lower in the DSD group compared with the acute illness group (mean differencesemean: 14.8 cm/s; 95% CI 8.0e21.6; P < .001; systolic: 29.0; 95% CI 16.0e41.9; P < .001 and diastolic: 8.9; 95% CI 2.5e15.3; P ¼ .003). The DSD group was also significantly lower on mean and diastolic FV than the AD (mean: 13.1 95% CI 0.7e25.5; P ¼ .040 and diastolic: 10.9;
G.A. Caplan et al. / JAMDA 15 (2014) 355e360
357
Table 1 Baseline Characteristics Characteristics
Alzheimer’s Dementia (n ¼ 10)
Delirium (n ¼ 12)
Delirium Superimposed on Dementia (n ¼ 8)
Acute Illness (n ¼ 14)
Age (mean SD) Sex (F:M) Number (%), Reason for admission
74 9 3:7 10 (100) Outpatient follow-up
84 8* 6:6 2 (17) Fall 4 (33) Infections: 1 (8) RTI, 2 (17) UTI, 1 (8) Other infections 1 (8) Delirium 5 (42) Other conditions
81 7 3:5 1 (12) Syncope 7 (87) Infections: 3 (37) RTI 4 (50) UTI
84 9* 2:12 5 (35) Fall 4 (28) Infections: 3 (21) RTI 1 (7) UTI 0 (0) Other infections 5 (35) Other conditions
9 1 6 3 0 1 3 3 0 2 2 3
3 0 1 2 8 1 3 2 0 1 4 2
9 3 1 4 0 1 8 6 0 0 3 0
Previous medical history, Number (%) Hypertension 3 (30) Congestive Cardiac failure 0 (0) Diabetes 0 (0) Cerebrovascular Accident 2 (20) Dementia 10 (100) Head injury 0 (0) Atrial Fibrillation 2 (20) Ischemic heart Disease 1 (10) Carotid artery Disease 0 (0) Current smoker 1 (10) Daily alcohol 2 (20) 2 (20) Use of psychotropic agent prior to and during hospitalization *P y P z P x P
< < < <
(75) (8) (50) (25) (0)y (8) (25) (25) (0) (17) (17) (25)
(38) (0) (13) (25) (100)z (12) (38) (25) (0) (12.5) (50) (25)
5 (26)
(64) (21) (7) (29) (0)y,x (7) (57) (43) (0) (0) (21) (0)
0 (0)
.05 v Alzheimer’s dementia. .001 v Alzheimer’s dementia. .001 v delirium. .001 v delirium þ pre-existing dementia.
95% CI 0.9e20.9; P ¼ .035) and delirium groups (mean: 9.4; 95% CI 2.7-16.3; P ¼ .009 and diastolic: 10.8 95% CI 4.2e17.4; P ¼ .001) (Figure 3). The PI was higher in the DSD group than in the AD or delirium only groups (mean differences: 0.6; 95% CI 0.2e0.9; P ¼ .002 and 0.6; 95% CI 0.3e0.8; P < .001) (Table 2, Figure 4). TCD was able to distinguish DSD from all other groups. A mean middle cerebral artery blood FV cut-off of 32.25 cm/s accurately
diagnoses delirium superimposed on dementia with a sensitivity of 0.875 and specificity of 0.788, area under the curve of 0.884; P ¼ .001 (Figure 2, online only). On multivariate analysis using logistic regression, after adjustment for age, initial mean FV remained significantly different between the DSD and other groups (adjusted odds ratio 0.762; 95% CI 0.614e0.948; P ¼ .013).
Table 2 Disease Measures and Transcranial Doppler Results Alzheimer’s Dementia (n ¼ 10)
Delirium (n ¼ 12)
CCIzz APACHE IIzz MMSE CAMzz Delirium indexzz GDSzz IQCODEzz Barthel index IADL Transcranial Doppler FV mean (cm/s) FV systolic (cm/s) FV diastolic (cm/s) PI
Delirium Superimposed on Dementia (n ¼ 8)
Acute Illness (n ¼ 14)
Mean Standard Deviation
Assessment Score 4.9 26.3 21.8 1.2 2.5 2.5 4.2 19.5 9.4
1.8 9.6 3.3 0.6 1.1 2.4 0.4 0.8 1.8
5.8 41.6 15.5 5.8 7.5 4.2 4.3 17.2 7.3
41.3 71.9 15.3 1.5
15.7 26.9 12.7 0.3
37.7 71.8 15.3 1.5
1.8* 10.6y 8.8* 1.2z 3.1z 3.7 0.6 3.6 3.7
7.4 44.6 11.3 5.9 9.4 1.7 4.6 15.1 5.8
1.6y 16.3y 8.4y 1.0z 3.9z 3.1 0.7 5.1* 3.4*
6.8 40.9 24.4 0.7 1.7 5.5 3.1 17.1 8.3
2.5 10.9y 5.2k,yy 0.7{,yy 1.5{,yy 2.9*,** 0.4z,{,yy 3.1* 3.0
8.2 13.8 8.2 0.3
28.2 61.6 4.4 2.1
4.7*,k 11.8 4.0*,k 0.3y,{
43.0 90.6 13.3 1.8
8.3yy 14.8k,yy 8.0** 0.4x
APACHE II, Acute Physiology and Chronic Health Evaluation II; CAM, Confusion Assessment Method; CCI, Charlson comorbidity index; FV, flow velocity; GDS, Geriatric Depression Scale; IQCODE, Informant Questionnaire on Cognitive Decline in the Elderly; IADL, Instrumental Activities of Daily Living; MMSE, Mini-Mental State Examination; PI, pulsatility index. *P < .05 vs Alzheimer’s dementia. y P < .01 vs Alzheimer’s dementia. z P < .001 vs Alzheimer’s dementia. x P < .05 vs delirium. k P < .01 vs delirium. { P < .001 vs delirium. **P < .01 vs delirium superimposed on dementia. yy P < .001 vs delirium superimposed on dementia. zz High score indicates greater impairment.
358
G.A. Caplan et al. / JAMDA 15 (2014) 355e360
Fig. 3. Boxplot showing FV according to group. Mean FV was lower in delirium with pre-existing AD than acute illness (P < .001) AD (P ¼ .040) and delirium (P ¼ .009) patients using Student t test. There was no significant difference in mean FV between delirium and AD groups and between AD and acute illness groups. AD, Alzheimer’s disease; FV, flow velocity; MCA, middle cerebral artery.
Looking at both delirium groups together, mean FV increased post resolution of delirium, assessed clinically and with the CAM, compared with baseline (difference 4.8 cm/s; 95% CI 1.8e7.8; P ¼ .006). There was no similar change seen in the acute illness group following resolution of their acute illness. However, there was no change in PI postdelirium compared with during delirium (difference 0.085; 95% CI e0.224 to 0.395; P ¼ .55). Multinomial logistic regression analysis examining the effect on mean FV using either a diagnosis of dementia (c2 45.3; P ¼ .047) and diagnosis of delirium (c2 50.9; P ¼ .014), or using the severity measures IQCODE (c2 ¼ 78.7; P < .001) and DI (c2 56.2; P ¼ .004), revealed that both dementia and delirium had independent significant influence on mean FV. Examining the whole population, Pearson’s correlation coefficient showed negative correlations between FV and PI, (R ¼ 0.5; P ¼ .001) as expected; FV and DI, (R ¼ 0.39; P ¼ .009) (Figure 5); FV and CAM, (R ¼ 0.39; P ¼ .010); and between FV and IQCODE, (R ¼ 0.41;
Fig. 4. Boxplot showing distribution of pulsatility index in the groups. The pulsatility index was higher in delirium superimposed on dementia than in AD or delirium only (P ¼ 0.002 and P < .001), but not than acute illness. AD, Alzheimer’s disease.
Fig. 5. MCA flow velocity and delirium index. Pearson correlation coefficient: R ¼ 0.39; P ¼ .009. DSD, delirium superimposed on dementia; MCA, middle cerebral artery.
P ¼ .005). Figure 5 illustrates that the worse FV in the DSD group is not simply because they had more severe delirium, and Table 2 shows that their mean DI was not worse than the delirium without dementia group. There was a positive correlation between FV and MMSE score, (R ¼ 0.33; P ¼.029). However, there was no significant correlation between FV and age, sex, APACHE II, CCI, GDS, Barthel index, and modified IADL score. The PI correlates with the CCI (R¼ 0.472; P ¼ .036). Discussion This is the first study to be able to diagnose DSD with a physiological measure, and the first showing a reversible reduction in CBF during delirium using a noninvasive method, TCD, as well as the first study to compare CBF in delirium and AD, and, thus, show that preexisting dementia is associated with worse CBF during delirium. CBF was significantly reduced during delirium and improved postdelirium, interestingly to a level close to that seen in the well AD group. The lack of similar variation in CBF in the acute illness group during their hospitalization, suggests that delirium itself accounted for temporary changes in CBF, or that decreased CBF triggered the delirium. The transcranial Doppler, which is noninvasive and nontoxic, showed high sensitivity and specificity for DSD making it potentially a useful clinical tool, although it is operator-dependent. Clinicians are often reluctant to diagnose DSD, having difficulty identifying the combination.10,11 This may be what causes a higher mortality for DSD12 as underlying medical problems may be missed. Although a higher mortality for DSD is not a universal finding,32 it additionally puts more burden on health systems because of subsequent higher readmission rates, possibly for untreated missed diagnoses.11 Thus, an objective test for DSD would be of great benefit for patients, clinicians, aged care facilities, and hospitals. Our finding of reduced CBF during delirium is similar to that seen in 14/19 studies in a recent review of neuroimaging in delirium, where 3 studies showed a mixed picture and 2 found only hyperperfusion.15 However, 7 of these studies were single patient case reports, and none of the studies included more participants than ours. The pattern of increased CBF after resolution of delirium in our study
G.A. Caplan et al. / JAMDA 15 (2014) 355e360
is also consistent with the increase seen on computed tomography CBF after delirium in a group of 10 younger patients with multiorgan failure or multiple injuries (mean age 47) in intensive care units (ICU),33 but in only 1/6 older (mean age 82) general medical patients, and only in parietal lobe perfusion, on single photon emission computed tomography scanning after delirium, whereas 2 showed a reduction in perfusion and three no change, although the delirium had not actually resolved by the second scan.34 Neither of these studies included controls, however the latter study did include a mixed group of older patients, similar to ours. Although CBF in our well group with AD is not significantly different to those with delirium without a history of dementia, an additive effect of DSD on CBF is seen, as indicated by the independently significant effects of dementia and delirium on CBF found in logistic regression. The correlation of CBF with measures of delirium such as the CAM and DI suggests that CBF plays a pathophysiological role in delirium, particularly when added to findings about glycolysis, because of the acute, transient nature of delirium symptomatology. Abnormal glucose metabolism, as detected by a fluorodeoxyglucose positron emission tomography scan, is the most sensitive and specific test for AD, and reduced CBF is also a feature of AD.14,19 Therefore, our findings of worse disturbances of glucose metabolism and CBF in delirium than AD, suggests that delirium should be viewed as an acute exacerbation of AD.8,13 Although in the pathophysiology of AD most attention falls on amyloid and tau protein deposition, these changes lack specificity, and are likely less useful in investigating the pathophysiological link with a transient condition like delirium than more dynamic characteristic changes like abnormal glucose metabolism and CBF. The exact mechanism for decreased CBF in delirium is unclear. It may be due to vascular smooth muscle damage or endothelial disturbance. ICU patients with sepsis and delirium showed impaired cerebral autoregulation, as indicated by an abnormal, positive correlation between mean arterial pressure and FV, although FV was not reduced.17 However, a second ICU study found preserved cerebral autoregulation in sepsis-associated delirium,35 while a third found impaired carbon dioxide-induced cerebral vasomotor reactivity in septic shock in patients who had some features of delirium.36 A fall in CBF could account for the disturbed cerebral glucose metabolism, manifested by increased cerebrospinal fluid lactate in delirium and decreased neuron specific enolase, suggesting a switch to anaerobic glycolysis.13 A slight fall in CBF might only increase extraction of oxygen and glucose by the brain from remaining blood flow, thus, allowing normal cerebral metabolism. A significant CBF fall may result in altered cerebral metabolism such as a switch to anaerobic glycolysis due to lack of oxygen. Altered cerebral metabolism can lead to widespread neuronal damage and cholinergic dysfunction, thus explaining the complex syndrome seen during delirium and the negative correlation between FV and DI. There are no other data comparing CBF changes between delirium and AD. Despite finding lower CBF in patients with DSD, compared with those with dementia alone, we did not find a significant difference between the overall groups with dementia and delirium. The evidence of MCA hypoperfusion predisposing to postoperative cognitive decline would be consistent with those patients having dementia, and supported by their apparently worse scores on cognitive tests preoperatively.18 We confirmed that PI and FV are strongly inversely correlated in older people, as they are in the general population, because the PI equals the difference between the systolic and diastolic velocities divided by the mean velocity. However, the relationship varied between groups, PI was lowest in the AD group, even though FV was highest in the acute illness group, suggesting disease-specific hemodynamic factors were at work. The interplay between CBF,
359
cerebral perfusion pressure, and cerebral autoregulation can lead to varied cerebral hemodynamic changes in different circumstances. Higher FV and PI in the acute illness group could be due to a disturbance of cerebral perfusion pressure in the presence of preserved autoregulation. In common with recent work, this study included patients with delirium from multiple etiologies to more accurately reflect the common clinical condition.13,34 Because delirium is a global process resulting in fluctuating mental status and widespread physiological disturbances, it is inherently challenging to study. Our study does have some limitations. First, there was a relatively small number of patients involved, although this is the largest study of CBF in delirium. Recruitment of patients with delirium or dementia is difficult and consent must generally be obtained from others. Also, lack of standardization of patient characteristics and the control of factors that could influence CBF and autoregulation such as PaCO2 across the groups could limit the reliability of the results obtained. Third, the “blind approach,” measuring FV without visualizing the MCA, taken might not correctly identify the MCA or standardize the portion of MCA being measured despite steps taken to ensure accuracy in detecting M1 portion of MCA, because of anatomic variation. A single comparison of FV readings between groups may not truly reflect blood flow changes attributable to delirium alone or give a conclusive finding on absolute CBF.17,37 This is due to the presence of various differences such as age and arterial partial pressure of carbon dioxide (PaCO2) that can affect CBF and the relationship between blood flow and FV. Hence, FV changes within subjects are a more accurate representation of CBF changes associated with delirium. Although FV was lower during delirium compared with postdelirium, the significance of any change in CBF remains uncertain as it could either be the cause or consequence of delirium, or an epiphenomenon. Also, a fall in FV might be due to impairment in cerebral autoregulation or variation in cerebral autoregulation rate. Hence, it may be helpful to investigate the role of cerebral autoregulation during delirium. TCD results could also be compared with other noninvasive methods such as TCD imaging, which allows more precise identification of MCA. More arterial segments comprising the circle of Willis could also be measured to give a more thorough evaluation of the circulatory changes during delirium. Conclusions The finding of a decreased FV in DSD and a reversible decrease in CBF during delirium, when added to findings about a switch to anaerobic glycolysis, provides a possible diagnostic test for DSD as well as important insights into the pathophysiology of delirium and a putative relationship with AD. TCD provides direct physical confirmation of brain perfusion impairment and might be a useful way to monitor delirium and even to assess the efficacy of medical interventions on CBF through dynamic monitoring. Because delirium is such an important clinical syndrome, and little is known about its pathophysiology, more detailed evaluation of cerebral hemodynamics in delirium would seem essential. Acknowledgments This study was assisted by staff from the Geriatric Department. The authors are thankful to Professor Christopher Levi, Stroke Unit, John Hunter Hospital, Dr Ramon Varcoe, Vascular Surgery Department, and staff of the Vascular Diagnostic Laboratory Prince of Wales Hospital and for their advice on the usage of TCD and the interpretation of some of the readings.
360
G.A. Caplan et al. / JAMDA 15 (2014) 355e360
Supplementary Data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.jamda.2013.12.079. References 1. Inouye SK. Delirium in older persons. N Engl J Med 2006;354:1157e1165. 2. Messinger-Rapport BJ, Morley JE, Thomas DR, Gammack JK. Clinical update on nursing home medicine: 2011. J Am Med Dir Assoc 2011;12:615e626. 3. Fong T, Tulebaev S, Inouye S. Delirium in elderly adults: Diagnosis, prevention and treatment. Nature Rev Neurol 2009;5:210e220. 4. Voyer P, Richard S, McCusker J, et al. Detection of delirium and its symptoms by nurses working in a long term care facility. J Am Med Dir Assoc 2012;13: 264e271. 5. Flaherty JH. Delirium. In: Pathy J, Sinclair A, Morley J, editors. Principles and Practice of Geriatric Medicine. 4th ed. Chichester: John Wiley and Sons; 2006. p. 1047e1060. 6. Thomas RI, Cameron DJ, Fahs MC. A prospective study of delirium and prolonged hospital stay. Arch Gen Psychiatr 1988;45:937e940. 7. Sands MB, Dantoc BP, Hartshorn A, et al. Single question in delirium (SQiD): Testing its efficacy against psychiatrist interview, the Confusion Assessment Method and the Memorial Delirium Assessment Scale. Palliat Med 2010;24: 561e565. 8. Fong TG, Jones RN, Shi P, et al. Delirium accelerates cognitive decline in Alzheimer disease. Neurology 2009;72:1570e1575. 9. Bellelli G, Speciale S, Morghen S, et al. Are fluctuations in motor performance a diagnostic sign of delirium? J Am Med Dir Assoc 2011;12:578e583. 10. Speed G, Wynaden D, McGowan S, et al. Prevalence rate of delirium at two hospitals in Western Australia. Aust J Adv Nurs 2007;25:38e43. 11. Fick D, Foreman M. Consequences of not recognizing delirium superimposed on dementia in hospitalized elderly individuals. J Gerontol Nurs 2000;26: 30e40. 12. Bellelli G, Frisoni GB, Turco R, et al. Delirium superimposed on dementia predicts 12-month survival in elderly patients discharged from a postacute rehabilitation facility. J Gerontol A Biol Sci Med Sci 2007;62A:1306e1309. 13. Caplan GA, Kvelde T, Lai C, et al. Cerebrospinal fluid in long-lasting delirium compared with Alzheimer’s dementia. J Gerontol A Biol Sci Med Sci 2010;65: 1130e1136. 14. Mosconi L, Pupi A, De Leon MJ. Brain glucose hypometabolism and oxidative stress in preclinical Alzheimer’s disease. Ann NY Acad Sci 2008;1147: 180e195. 15. Alsop D, Fearing M, Johnson K, et al. The role of neuroimaging in elucidating delirium pathophysiology. J Gerontol A Biol Sci Med Sci 2006;61: 1287e1293. 16. Soiza R, Sharma V, Ferguson K, et al. Neuroimaging studies of delirium: A systematic review. J Psychosomatic Res 2008;65:239e248. 17. Pfister D, Siegemund M, Dell-Kuster S, et al. Cerebral perfusion in sepsisassociated delirium. Crit Care 2008;2:R63eR71.
18. Benvenuti SM, Zanata P, Longo C, et al. Preoperative cerebral hypoperfusion in the left, not in the right, hemisphere is associated with cognitive decline after cardiac surgery. Psychosom Med 2012;74:73e80. 19. Sabayan B, Jansen S, Oleksik AM, et al. Cerebrovascular hemodynamics in Alzheimer’s disease and vascular dementia: A meta-analysis of transcranial Doppler studies. Ageing Res Rev 2012;11:271e277. 20. Marsden KR, Haykowsky MJ, Smirl JD, et al. Aging blunts hyperventilationinduced hypocapnia and reduction in cerebral blood flow velocity during maximal exercise. Age 2012;34:725e735. 21. Inouye SK, Van CH, Alessi CA, et al. Clarifying confusion: The confusion assessment method, a new method for detection of delirium. Ann Int Med 1990;113:941e948. 22. Folstein MF, Folstein SE, McHugh PR. Mini-Mental State: A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12: 189e198. 23. McCusker J, Cole MG, Bellavance F. Reliability and validity of a new measure of severity of delirium. Int Psychogeriatr 1998;10:421e433. 24. Jorm A. The informant questionnaire on cognitive decline in the elderly (IQCODE): A review. Int Psychogeriatr 2004;16:275e293. 25. Yesavage J, Brink TL, Rose TL, et al. Development and validation of a geriatric depression scale: A preliminary report. J Psychiatr Res 1983;17:37e49. 26. Mahoney FI, Barthel DW. Functional evaluation: The Barthel index. Maryland Med J 1965;14:61e65. 27. Lawton M, Moss M, Fulcomer M, Kleban MH. A research and service oriented multilevel assessment instrument. J Gerontol 1982;37:91e95. 28. Knaus W, Draper EA, Wagner DP, Zimmerman JE. APACHE II: A severity of disease classification system. Crit Care Med 1985;13:818e829. 29. Charlson M, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 1987;40:373e383. 30. Aaslid R, Markwalder T, Nornes H. Noninvasive transcranial doppler ultrasound recording of flow velocity in basal cerebral arteries. J Neurosurg 1982;57: 769e774. 31. Krejza J, Mariak Z, Melhem ER, Bert RJ. A guide to the identification of major cerebral arteries with transcranial color Doppler sonography. AJR 2000;174: 1297e1303. 32. McCusker J, Cole M, Abrahamovicz M, et al. Delirium predicts 12-month mortality. Arch Intern Med 2002;162:457e463. 33. Yokota H, Ogawa S, Kurokawa A, Yamamoto Y. Regional cerebral blood flow in delirium patients. Psychiatry Clin Neurosci 2003;57:337e339. 34. Fong TG, Bogardus ST, Daftary A, et al. Cerebral perfusion changes in older delirious patients using 99mTc HMPAO SPECT. J Gerontol A Biol Sci Med Sci 2006;61A:1294e1299. 35. Matta BF, Stow PJ. Sepsis-induced vasoparalysis does not involve the cerebral vasculature: Indirect evidence from autoregulation and carbon dioxide reactivity studies. Br J Anaesth 1996;76:790e794. 36. Terborg C, Schummer W, Albrecht M, et al. Dysfunction of vasomotor reactivity in severe sepsis and septic shock. Intensive Care Med 2001;27: 1231e1234. 37. Thees C, Kaiser M, Scholz M, et al. Cerebral haemodynamics and carbon dioxide reactivity during sepsis syndrome. Crit Care 2007;11:R123.
JAMDA journal homepage: www.jamda.com
Original Study
Transcranial Doppler to Measure Cerebral Blood Flow in Delirium Superimposed on Dementia. A Cohort Study Gideon A. Caplan MD a, b, *, ZhongZheng Lan BSc(Med) c, Lyndal Newton MBBS a, b, Tasha Kvelde DPsy a, Catherine McVeigh MD a, b, Mark A. Hill PhD c a
Department of Geriatric Medicine, Prince of Wales Hospital, Sydney, Australia Prince of Wales Clinical School, University of New South Wales, Sydney, Australia c School of Medical Sciences, University of New South Wales, Sydney, Australia b
a b s t r a c t Keywords: Delirium Alzheimer’s disease ultrasound assessment of cognitive disorders/dementia geriatric assessment
Objective: Delirium superimposed on dementia (DSD) is frequently not diagnosed, at great cost. Both delirium and dementia are associated with cerebral hypoperfusion. A switch to anaerobic glycolysis in the central nervous system during delirium compared to Alzheimer’s dementia (AD) suggests greater hypoperfusion in DSD. The main aims of this study were to investigate whether cerebral hypoperfusion could differentiate DSD from related entities, and the characteristics of that hypoperfusion. Methods: Prospective cohort study of 44 Geriatric Medicine patients in 4 groups: (1) delirium, no history of dementia; (2) DSD; (3) acute illness without delirium or dementia; and (4) AD, no delirium. We measured CBF using transcranial Doppler to assess flow velocity (FV) and pulsatility index in the middle cerebral artery (MCA). Results: DSD has lower FV than either AD or delirium alone, or acute illness (28.2 4.7 vs AD: 41.3 15.7; P ¼ .040; vs delirium 37.7 8.2; P ¼.009; vs acute illness 43.0 8.3; P <.001). A mean MCA FV cut-off of 32.25 cm/s diagnoses DSD with a sensitivity of 0.875 and specificity of 0.788, area under the curve 0.884; P ¼ .001. Resolution of delirium improves FV (P ¼ .005). FV correlates with delirium severity (delirium index R ¼ 0.39; P ¼ .009) and dementia (Mini-Mental State Examination, R ¼ 0.33; P ¼ .029, and Informant Questionnaire on Cognitive Decline in the Elderly, R ¼ 0.41; P ¼ .005). Conclusions: Transcranial Doppler is a potential diagnostic and monitoring test for DSD. Correlation with clinical indicators of delirium suggests pathophysiological significance. Crown Copyright Ó 2014 Published by Elsevier Inc. on behalf of American Medical Directors Association, Inc. All rights reserved.
Delirium is a serious complication of illness, commonly affecting older patients, including more than 20% of hospitalized people over 65, and 70%e87% of those in intensive care, as well as being a major problem in long-term care.1,2 It is a significant cause of preventable morbidity and mortality in older patients, substantially increasing healthcare utilization and costs.1,3 Diagnosis is clinical, importantly informed by the corroborative history, however, 85% of delirious patients remain undiagnosed in hospital and 50% in nursing homes, or are misdiagnosed as having dementia, despite delirium being a medical emergency.1,4 This is partly because of a lack of diagnostic This study was supported by an unrestricted grant from the Julia Lowy Foundation. Z.Z.L. was partially funded by a scholarship from the NSW Branch of the Australian and New Zealand Society for Geriatric Medicine. The authors declare no conflicts of interest. * Address correspondence to Gideon A. Caplan, MD, Department of Geriatric Medicine, Prince of Wales Hospital, Edmund Blackett Building, Randwick, New South Wales 2031, Australia. E-mail address: [email protected] (G.A. Caplan).
tests, especially when superimposed on dementia.5,6 Diagnostic indices are available, such as the Confusion Assessment Method (CAM), but it is unreliable in untrained hands.7 Delirium is more common in people with dementia, is the most common acute driver of dementia, with which it shares many risk factors, and causes worsening of cognition and performance of activities of daily living, the clinical measures of dementia.8,9 Delirium superimposed on dementia (DSD) is more frequently missed and results in higher mortality and readmission rates.10e12 We recently showed that cerebrospinal fluid lactate is elevated and neurone specific enolase decreased in delirium suggesting a switch to anaerobic glycolysis,13 compared with Alzheimer’s dementia (AD) where altered glucose metabolism is the most sensitive and specific marker of AD.14 Abnormal glycolysis could lead to widespread neuronal dysfunction, triggering the clinical manifestations of delirium. Changes in glycolysis could be due to excessive demand for energy, toxins or reduced cerebral blood flow (CBF). The brain is dependent on glucose for energy production and inadequate
1525-8610/$ - see front matter Crown Copyright Ó 2014 Published by Elsevier Inc. on behalf of American Medical Directors Association, Inc. All rights reserved. http://dx.doi.org/10.1016/j.jamda.2013.12.079
356
G.A. Caplan et al. / JAMDA 15 (2014) 355e360
CBF could disrupt glucose transport or its complete oxidation, thus lowering the efficiency of glucose metabolism.13 Studies of CBF in delirium have been inconclusive, with variable findings ranging from no change to hypoperfusion and hyperperfusion.15e17 Cerebral hypoperfusion predisposes to postoperative cognitive decline, which is linked to delirium.18 Many patients with delirium also have dementia so the issue is complicated by evidence that decreased CBF is characteristic of AD, which also predisposes to delirium. However, CBF in delirium and AD has never been compared, therefore, it is unclear if the changes seen in delirium are due to underlying AD.19 In addition, CBF declines by between 28% and 50% from age 30 to age 70.20 We investigated the hypothesis that DSD is associated with greater reductions in CBF by noninvasive, safe transcranial Doppler (TCD). Methods This cohort study was approved by the Human Research Ethic Committee. Written informed consent was obtained from all patients or their person responsible, where the patient lacked capacity. In 2011 we recruited consecutive consenting patients in four groups with (1) acute illness and delirium but no history of dementia; (2) acute illness with delirium and a history of dementia, the DSD group; (3) acute illness without delirium or dementia; and (4) no acute illness or delirium, known (Alzheimer’s dementia) AD. Groups 1e3 were recruited from the Acute Geriatrics inpatient ward and were diagnosed by geriatricians based on clinical experience and the CAM.21 Patients admitted to the Geriatrics ward are screened with the CAM in the Emergency Department. Patients with AD were recruited from both a general geriatric outpatient clinic and a cognitive disorders clinic, with AD diagnosis based on Diagnostic and Statistical Manual, 4th edition, and National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association criteria for probable Alzheimer’s dementia. Cognitive and psychological assessments were conducted by a geriatrician and a neuropsychologist. Patients were excluded if they had (1) possible impaired cerebral vasoreactivity because of (a) sonographic evidence of severe extracranial stenosis (>70%); (b) sonographic evidence of intracranial stenosis, and (c) presence of territorial infarcts larger than one third of the hemisphere; and (2) no insonation window or a poor signal. Patients with underlying dementia or history of delirium during current hospitalization were excluded from the acute illness group. Patients were first screened for the presence of an insonation window. After a TCD measurement was taken, all subjects underwent baseline assessments through the completion of the clinical measures listed below. No intervention was performed on any of the patients for this observational study. Assessments Cognitive and psychological assessments were conducted using the Mini-Mental State Examination (MMSE);22 the Confusion Assessment Method (CAM);21 the delirium index (DI);23 the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE);24 and the Geriatric Depression Scale (GDS).25 Physiological and functional status assessment were carried out using the Barthel index;26 the modified Instrumental Activities of Daily Living (IADL) index;27 the Acute Physiology, Age, Chronic Health Evaluation (APACHE) II;28 and the Charlson comorbidity index (CCI).29 The IQCODE was asked in regard to the patient’s status relative to the pre-hospitalization status. For patients in groups 1e3 who were in hospital, the Doppler and CAM were repeated second daily until discharge. No testing was done prior to hospitalization. Group 4 were assessed only once. The order of testing was the same for all groups. The results in Table 2 refer to the initial testing.
TCD Imaging We used the QL modular software and Doppler-Box from DWL Compumedics for analyzing transcranial spectral signals derived from systolic velocity, diastolic velocity, FV, and pulsatility indices generated by the MCA. The transtemporal approach of insonation was performed with the patient either lying supine on the examination bed or sitting upright in a chair and the head straight in both positions. The 2 MHz pulsed TCD probe was positioned according to the criteria of Aaslid,30,31 which allowed identification of the MCA. All measurements were stored on hard disk for off-line analysis. The average of right and left FV and PI were used for analyses if both MCA could be insonated adequately. A single measurement was made on every second day of their hospitalization for patients in the delirium and acute illness groups. A one-off reading was taken from all patients with AD from the outpatient clinic. All TCD readings were taken by one person (Z.Z.L.) to reduce variability. Statistical Analysis SPSS 21 for Windows (SPSS Inc, Chicago, IL) was used in this study. A 2-tailed P value of less than .05 was considered significant. For parametric data, t tests were used for univariate analysis, while Mann-Whitney U was used for nonparametric data. Multivariate analysis was performed with logistic and binomial regression. Fisher exact test was used to compare proportions. Receiver operating characteristics curve was calculated using a nonparametric distribution assumption. Missing data were excluded listwise. A sample size calculation indicated that this sample size would be sufficient to detect a 20% difference in CBF. Groups were compared on their initial TCD and CAM only. However, initial and postdelirium TCD was compared in groups 1þ2 with initial and final TCD in group 3. Results Between May and September 2011, a total of 86 patients were screened and 44 patients were eligible for inclusion and consented to participate in this study (Figure 1, online only). Patient characteristics (Table 1) were similar in the 3 groups except for age. Age was significantly lower in the AD group compared with the delirium [difference 9.5 years; 95% confidence interval (CI) 1.4e17.7; P ¼ .025], and acute illness groups (10.0 years; 95% CI 2.1e18.0; P ¼ .026). There were no significant differences in the reasons for hospitalization between the acute illness and delirium groups. There was a trend toward more patients with diabetes and more treated with psychotropics in the delirium group. There were significant differences among the 3 groups in the following measures: APACHE II, MMSE, CAM, DI, Barthel index, IQCODE, and modified IADL according to the expected clinical characteristics of each group (Table 2). However, the raised IQCODE in the delirium without dementia group may indicate that respondents had difficulty separating the cognitive decline associated with delirium from the premorbid state. The delirium and acute illness groups had similar acuity of illness on the APACHE II. The dementia and delirium groups had similar cognitive decline on the IQCODE, although the MMSE was lower in the delirium groups than in AD and acute illness, cognitive decline being a feature of both. The delirium groups scored more highly on both measures of delirium, the CAM and DI, than other groups. On univariate analysis mean, systolic and diastolic FV was lower in the DSD group compared with the acute illness group (mean differencesemean: 14.8 cm/s; 95% CI 8.0e21.6; P < .001; systolic: 29.0; 95% CI 16.0e41.9; P < .001 and diastolic: 8.9; 95% CI 2.5e15.3; P ¼ .003). The DSD group was also significantly lower on mean and diastolic FV than the AD (mean: 13.1 95% CI 0.7e25.5; P ¼ .040 and diastolic: 10.9;
G.A. Caplan et al. / JAMDA 15 (2014) 355e360
357
Table 1 Baseline Characteristics Characteristics
Alzheimer’s Dementia (n ¼ 10)
Delirium (n ¼ 12)
Delirium Superimposed on Dementia (n ¼ 8)
Acute Illness (n ¼ 14)
Age (mean SD) Sex (F:M) Number (%), Reason for admission
74 9 3:7 10 (100) Outpatient follow-up
84 8* 6:6 2 (17) Fall 4 (33) Infections: 1 (8) RTI, 2 (17) UTI, 1 (8) Other infections 1 (8) Delirium 5 (42) Other conditions
81 7 3:5 1 (12) Syncope 7 (87) Infections: 3 (37) RTI 4 (50) UTI
84 9* 2:12 5 (35) Fall 4 (28) Infections: 3 (21) RTI 1 (7) UTI 0 (0) Other infections 5 (35) Other conditions
9 1 6 3 0 1 3 3 0 2 2 3
3 0 1 2 8 1 3 2 0 1 4 2
9 3 1 4 0 1 8 6 0 0 3 0
Previous medical history, Number (%) Hypertension 3 (30) Congestive Cardiac failure 0 (0) Diabetes 0 (0) Cerebrovascular Accident 2 (20) Dementia 10 (100) Head injury 0 (0) Atrial Fibrillation 2 (20) Ischemic heart Disease 1 (10) Carotid artery Disease 0 (0) Current smoker 1 (10) Daily alcohol 2 (20) 2 (20) Use of psychotropic agent prior to and during hospitalization *P y P z P x P
< < < <
(75) (8) (50) (25) (0)y (8) (25) (25) (0) (17) (17) (25)
(38) (0) (13) (25) (100)z (12) (38) (25) (0) (12.5) (50) (25)
5 (26)
(64) (21) (7) (29) (0)y,x (7) (57) (43) (0) (0) (21) (0)
0 (0)
.05 v Alzheimer’s dementia. .001 v Alzheimer’s dementia. .001 v delirium. .001 v delirium þ pre-existing dementia.
95% CI 0.9e20.9; P ¼ .035) and delirium groups (mean: 9.4; 95% CI 2.7-16.3; P ¼ .009 and diastolic: 10.8 95% CI 4.2e17.4; P ¼ .001) (Figure 3). The PI was higher in the DSD group than in the AD or delirium only groups (mean differences: 0.6; 95% CI 0.2e0.9; P ¼ .002 and 0.6; 95% CI 0.3e0.8; P < .001) (Table 2, Figure 4). TCD was able to distinguish DSD from all other groups. A mean middle cerebral artery blood FV cut-off of 32.25 cm/s accurately
diagnoses delirium superimposed on dementia with a sensitivity of 0.875 and specificity of 0.788, area under the curve of 0.884; P ¼ .001 (Figure 2, online only). On multivariate analysis using logistic regression, after adjustment for age, initial mean FV remained significantly different between the DSD and other groups (adjusted odds ratio 0.762; 95% CI 0.614e0.948; P ¼ .013).
Table 2 Disease Measures and Transcranial Doppler Results Alzheimer’s Dementia (n ¼ 10)
Delirium (n ¼ 12)
CCIzz APACHE IIzz MMSE CAMzz Delirium indexzz GDSzz IQCODEzz Barthel index IADL Transcranial Doppler FV mean (cm/s) FV systolic (cm/s) FV diastolic (cm/s) PI
Delirium Superimposed on Dementia (n ¼ 8)
Acute Illness (n ¼ 14)
Mean Standard Deviation
Assessment Score 4.9 26.3 21.8 1.2 2.5 2.5 4.2 19.5 9.4
1.8 9.6 3.3 0.6 1.1 2.4 0.4 0.8 1.8
5.8 41.6 15.5 5.8 7.5 4.2 4.3 17.2 7.3
41.3 71.9 15.3 1.5
15.7 26.9 12.7 0.3
37.7 71.8 15.3 1.5
1.8* 10.6y 8.8* 1.2z 3.1z 3.7 0.6 3.6 3.7
7.4 44.6 11.3 5.9 9.4 1.7 4.6 15.1 5.8
1.6y 16.3y 8.4y 1.0z 3.9z 3.1 0.7 5.1* 3.4*
6.8 40.9 24.4 0.7 1.7 5.5 3.1 17.1 8.3
2.5 10.9y 5.2k,yy 0.7{,yy 1.5{,yy 2.9*,** 0.4z,{,yy 3.1* 3.0
8.2 13.8 8.2 0.3
28.2 61.6 4.4 2.1
4.7*,k 11.8 4.0*,k 0.3y,{
43.0 90.6 13.3 1.8
8.3yy 14.8k,yy 8.0** 0.4x
APACHE II, Acute Physiology and Chronic Health Evaluation II; CAM, Confusion Assessment Method; CCI, Charlson comorbidity index; FV, flow velocity; GDS, Geriatric Depression Scale; IQCODE, Informant Questionnaire on Cognitive Decline in the Elderly; IADL, Instrumental Activities of Daily Living; MMSE, Mini-Mental State Examination; PI, pulsatility index. *P < .05 vs Alzheimer’s dementia. y P < .01 vs Alzheimer’s dementia. z P < .001 vs Alzheimer’s dementia. x P < .05 vs delirium. k P < .01 vs delirium. { P < .001 vs delirium. **P < .01 vs delirium superimposed on dementia. yy P < .001 vs delirium superimposed on dementia. zz High score indicates greater impairment.
358
G.A. Caplan et al. / JAMDA 15 (2014) 355e360
Fig. 3. Boxplot showing FV according to group. Mean FV was lower in delirium with pre-existing AD than acute illness (P < .001) AD (P ¼ .040) and delirium (P ¼ .009) patients using Student t test. There was no significant difference in mean FV between delirium and AD groups and between AD and acute illness groups. AD, Alzheimer’s disease; FV, flow velocity; MCA, middle cerebral artery.
Looking at both delirium groups together, mean FV increased post resolution of delirium, assessed clinically and with the CAM, compared with baseline (difference 4.8 cm/s; 95% CI 1.8e7.8; P ¼ .006). There was no similar change seen in the acute illness group following resolution of their acute illness. However, there was no change in PI postdelirium compared with during delirium (difference 0.085; 95% CI e0.224 to 0.395; P ¼ .55). Multinomial logistic regression analysis examining the effect on mean FV using either a diagnosis of dementia (c2 45.3; P ¼ .047) and diagnosis of delirium (c2 50.9; P ¼ .014), or using the severity measures IQCODE (c2 ¼ 78.7; P < .001) and DI (c2 56.2; P ¼ .004), revealed that both dementia and delirium had independent significant influence on mean FV. Examining the whole population, Pearson’s correlation coefficient showed negative correlations between FV and PI, (R ¼ 0.5; P ¼ .001) as expected; FV and DI, (R ¼ 0.39; P ¼ .009) (Figure 5); FV and CAM, (R ¼ 0.39; P ¼ .010); and between FV and IQCODE, (R ¼ 0.41;
Fig. 4. Boxplot showing distribution of pulsatility index in the groups. The pulsatility index was higher in delirium superimposed on dementia than in AD or delirium only (P ¼ 0.002 and P < .001), but not than acute illness. AD, Alzheimer’s disease.
Fig. 5. MCA flow velocity and delirium index. Pearson correlation coefficient: R ¼ 0.39; P ¼ .009. DSD, delirium superimposed on dementia; MCA, middle cerebral artery.
P ¼ .005). Figure 5 illustrates that the worse FV in the DSD group is not simply because they had more severe delirium, and Table 2 shows that their mean DI was not worse than the delirium without dementia group. There was a positive correlation between FV and MMSE score, (R ¼ 0.33; P ¼.029). However, there was no significant correlation between FV and age, sex, APACHE II, CCI, GDS, Barthel index, and modified IADL score. The PI correlates with the CCI (R¼ 0.472; P ¼ .036). Discussion This is the first study to be able to diagnose DSD with a physiological measure, and the first showing a reversible reduction in CBF during delirium using a noninvasive method, TCD, as well as the first study to compare CBF in delirium and AD, and, thus, show that preexisting dementia is associated with worse CBF during delirium. CBF was significantly reduced during delirium and improved postdelirium, interestingly to a level close to that seen in the well AD group. The lack of similar variation in CBF in the acute illness group during their hospitalization, suggests that delirium itself accounted for temporary changes in CBF, or that decreased CBF triggered the delirium. The transcranial Doppler, which is noninvasive and nontoxic, showed high sensitivity and specificity for DSD making it potentially a useful clinical tool, although it is operator-dependent. Clinicians are often reluctant to diagnose DSD, having difficulty identifying the combination.10,11 This may be what causes a higher mortality for DSD12 as underlying medical problems may be missed. Although a higher mortality for DSD is not a universal finding,32 it additionally puts more burden on health systems because of subsequent higher readmission rates, possibly for untreated missed diagnoses.11 Thus, an objective test for DSD would be of great benefit for patients, clinicians, aged care facilities, and hospitals. Our finding of reduced CBF during delirium is similar to that seen in 14/19 studies in a recent review of neuroimaging in delirium, where 3 studies showed a mixed picture and 2 found only hyperperfusion.15 However, 7 of these studies were single patient case reports, and none of the studies included more participants than ours. The pattern of increased CBF after resolution of delirium in our study
G.A. Caplan et al. / JAMDA 15 (2014) 355e360
is also consistent with the increase seen on computed tomography CBF after delirium in a group of 10 younger patients with multiorgan failure or multiple injuries (mean age 47) in intensive care units (ICU),33 but in only 1/6 older (mean age 82) general medical patients, and only in parietal lobe perfusion, on single photon emission computed tomography scanning after delirium, whereas 2 showed a reduction in perfusion and three no change, although the delirium had not actually resolved by the second scan.34 Neither of these studies included controls, however the latter study did include a mixed group of older patients, similar to ours. Although CBF in our well group with AD is not significantly different to those with delirium without a history of dementia, an additive effect of DSD on CBF is seen, as indicated by the independently significant effects of dementia and delirium on CBF found in logistic regression. The correlation of CBF with measures of delirium such as the CAM and DI suggests that CBF plays a pathophysiological role in delirium, particularly when added to findings about glycolysis, because of the acute, transient nature of delirium symptomatology. Abnormal glucose metabolism, as detected by a fluorodeoxyglucose positron emission tomography scan, is the most sensitive and specific test for AD, and reduced CBF is also a feature of AD.14,19 Therefore, our findings of worse disturbances of glucose metabolism and CBF in delirium than AD, suggests that delirium should be viewed as an acute exacerbation of AD.8,13 Although in the pathophysiology of AD most attention falls on amyloid and tau protein deposition, these changes lack specificity, and are likely less useful in investigating the pathophysiological link with a transient condition like delirium than more dynamic characteristic changes like abnormal glucose metabolism and CBF. The exact mechanism for decreased CBF in delirium is unclear. It may be due to vascular smooth muscle damage or endothelial disturbance. ICU patients with sepsis and delirium showed impaired cerebral autoregulation, as indicated by an abnormal, positive correlation between mean arterial pressure and FV, although FV was not reduced.17 However, a second ICU study found preserved cerebral autoregulation in sepsis-associated delirium,35 while a third found impaired carbon dioxide-induced cerebral vasomotor reactivity in septic shock in patients who had some features of delirium.36 A fall in CBF could account for the disturbed cerebral glucose metabolism, manifested by increased cerebrospinal fluid lactate in delirium and decreased neuron specific enolase, suggesting a switch to anaerobic glycolysis.13 A slight fall in CBF might only increase extraction of oxygen and glucose by the brain from remaining blood flow, thus, allowing normal cerebral metabolism. A significant CBF fall may result in altered cerebral metabolism such as a switch to anaerobic glycolysis due to lack of oxygen. Altered cerebral metabolism can lead to widespread neuronal damage and cholinergic dysfunction, thus explaining the complex syndrome seen during delirium and the negative correlation between FV and DI. There are no other data comparing CBF changes between delirium and AD. Despite finding lower CBF in patients with DSD, compared with those with dementia alone, we did not find a significant difference between the overall groups with dementia and delirium. The evidence of MCA hypoperfusion predisposing to postoperative cognitive decline would be consistent with those patients having dementia, and supported by their apparently worse scores on cognitive tests preoperatively.18 We confirmed that PI and FV are strongly inversely correlated in older people, as they are in the general population, because the PI equals the difference between the systolic and diastolic velocities divided by the mean velocity. However, the relationship varied between groups, PI was lowest in the AD group, even though FV was highest in the acute illness group, suggesting disease-specific hemodynamic factors were at work. The interplay between CBF,
359
cerebral perfusion pressure, and cerebral autoregulation can lead to varied cerebral hemodynamic changes in different circumstances. Higher FV and PI in the acute illness group could be due to a disturbance of cerebral perfusion pressure in the presence of preserved autoregulation. In common with recent work, this study included patients with delirium from multiple etiologies to more accurately reflect the common clinical condition.13,34 Because delirium is a global process resulting in fluctuating mental status and widespread physiological disturbances, it is inherently challenging to study. Our study does have some limitations. First, there was a relatively small number of patients involved, although this is the largest study of CBF in delirium. Recruitment of patients with delirium or dementia is difficult and consent must generally be obtained from others. Also, lack of standardization of patient characteristics and the control of factors that could influence CBF and autoregulation such as PaCO2 across the groups could limit the reliability of the results obtained. Third, the “blind approach,” measuring FV without visualizing the MCA, taken might not correctly identify the MCA or standardize the portion of MCA being measured despite steps taken to ensure accuracy in detecting M1 portion of MCA, because of anatomic variation. A single comparison of FV readings between groups may not truly reflect blood flow changes attributable to delirium alone or give a conclusive finding on absolute CBF.17,37 This is due to the presence of various differences such as age and arterial partial pressure of carbon dioxide (PaCO2) that can affect CBF and the relationship between blood flow and FV. Hence, FV changes within subjects are a more accurate representation of CBF changes associated with delirium. Although FV was lower during delirium compared with postdelirium, the significance of any change in CBF remains uncertain as it could either be the cause or consequence of delirium, or an epiphenomenon. Also, a fall in FV might be due to impairment in cerebral autoregulation or variation in cerebral autoregulation rate. Hence, it may be helpful to investigate the role of cerebral autoregulation during delirium. TCD results could also be compared with other noninvasive methods such as TCD imaging, which allows more precise identification of MCA. More arterial segments comprising the circle of Willis could also be measured to give a more thorough evaluation of the circulatory changes during delirium. Conclusions The finding of a decreased FV in DSD and a reversible decrease in CBF during delirium, when added to findings about a switch to anaerobic glycolysis, provides a possible diagnostic test for DSD as well as important insights into the pathophysiology of delirium and a putative relationship with AD. TCD provides direct physical confirmation of brain perfusion impairment and might be a useful way to monitor delirium and even to assess the efficacy of medical interventions on CBF through dynamic monitoring. Because delirium is such an important clinical syndrome, and little is known about its pathophysiology, more detailed evaluation of cerebral hemodynamics in delirium would seem essential. Acknowledgments This study was assisted by staff from the Geriatric Department. The authors are thankful to Professor Christopher Levi, Stroke Unit, John Hunter Hospital, Dr Ramon Varcoe, Vascular Surgery Department, and staff of the Vascular Diagnostic Laboratory Prince of Wales Hospital and for their advice on the usage of TCD and the interpretation of some of the readings.
360
G.A. Caplan et al. / JAMDA 15 (2014) 355e360
Supplementary Data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.jamda.2013.12.079. References 1. Inouye SK. Delirium in older persons. N Engl J Med 2006;354:1157e1165. 2. Messinger-Rapport BJ, Morley JE, Thomas DR, Gammack JK. Clinical update on nursing home medicine: 2011. J Am Med Dir Assoc 2011;12:615e626. 3. Fong T, Tulebaev S, Inouye S. Delirium in elderly adults: Diagnosis, prevention and treatment. Nature Rev Neurol 2009;5:210e220. 4. Voyer P, Richard S, McCusker J, et al. Detection of delirium and its symptoms by nurses working in a long term care facility. J Am Med Dir Assoc 2012;13: 264e271. 5. Flaherty JH. Delirium. In: Pathy J, Sinclair A, Morley J, editors. Principles and Practice of Geriatric Medicine. 4th ed. Chichester: John Wiley and Sons; 2006. p. 1047e1060. 6. Thomas RI, Cameron DJ, Fahs MC. A prospective study of delirium and prolonged hospital stay. Arch Gen Psychiatr 1988;45:937e940. 7. Sands MB, Dantoc BP, Hartshorn A, et al. Single question in delirium (SQiD): Testing its efficacy against psychiatrist interview, the Confusion Assessment Method and the Memorial Delirium Assessment Scale. Palliat Med 2010;24: 561e565. 8. Fong TG, Jones RN, Shi P, et al. Delirium accelerates cognitive decline in Alzheimer disease. Neurology 2009;72:1570e1575. 9. Bellelli G, Speciale S, Morghen S, et al. Are fluctuations in motor performance a diagnostic sign of delirium? J Am Med Dir Assoc 2011;12:578e583. 10. Speed G, Wynaden D, McGowan S, et al. Prevalence rate of delirium at two hospitals in Western Australia. Aust J Adv Nurs 2007;25:38e43. 11. Fick D, Foreman M. Consequences of not recognizing delirium superimposed on dementia in hospitalized elderly individuals. J Gerontol Nurs 2000;26: 30e40. 12. Bellelli G, Frisoni GB, Turco R, et al. Delirium superimposed on dementia predicts 12-month survival in elderly patients discharged from a postacute rehabilitation facility. J Gerontol A Biol Sci Med Sci 2007;62A:1306e1309. 13. Caplan GA, Kvelde T, Lai C, et al. Cerebrospinal fluid in long-lasting delirium compared with Alzheimer’s dementia. J Gerontol A Biol Sci Med Sci 2010;65: 1130e1136. 14. Mosconi L, Pupi A, De Leon MJ. Brain glucose hypometabolism and oxidative stress in preclinical Alzheimer’s disease. Ann NY Acad Sci 2008;1147: 180e195. 15. Alsop D, Fearing M, Johnson K, et al. The role of neuroimaging in elucidating delirium pathophysiology. J Gerontol A Biol Sci Med Sci 2006;61: 1287e1293. 16. Soiza R, Sharma V, Ferguson K, et al. Neuroimaging studies of delirium: A systematic review. J Psychosomatic Res 2008;65:239e248. 17. Pfister D, Siegemund M, Dell-Kuster S, et al. Cerebral perfusion in sepsisassociated delirium. Crit Care 2008;2:R63eR71.
18. Benvenuti SM, Zanata P, Longo C, et al. Preoperative cerebral hypoperfusion in the left, not in the right, hemisphere is associated with cognitive decline after cardiac surgery. Psychosom Med 2012;74:73e80. 19. Sabayan B, Jansen S, Oleksik AM, et al. Cerebrovascular hemodynamics in Alzheimer’s disease and vascular dementia: A meta-analysis of transcranial Doppler studies. Ageing Res Rev 2012;11:271e277. 20. Marsden KR, Haykowsky MJ, Smirl JD, et al. Aging blunts hyperventilationinduced hypocapnia and reduction in cerebral blood flow velocity during maximal exercise. Age 2012;34:725e735. 21. Inouye SK, Van CH, Alessi CA, et al. Clarifying confusion: The confusion assessment method, a new method for detection of delirium. Ann Int Med 1990;113:941e948. 22. Folstein MF, Folstein SE, McHugh PR. Mini-Mental State: A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12: 189e198. 23. McCusker J, Cole MG, Bellavance F. Reliability and validity of a new measure of severity of delirium. Int Psychogeriatr 1998;10:421e433. 24. Jorm A. The informant questionnaire on cognitive decline in the elderly (IQCODE): A review. Int Psychogeriatr 2004;16:275e293. 25. Yesavage J, Brink TL, Rose TL, et al. Development and validation of a geriatric depression scale: A preliminary report. J Psychiatr Res 1983;17:37e49. 26. Mahoney FI, Barthel DW. Functional evaluation: The Barthel index. Maryland Med J 1965;14:61e65. 27. Lawton M, Moss M, Fulcomer M, Kleban MH. A research and service oriented multilevel assessment instrument. J Gerontol 1982;37:91e95. 28. Knaus W, Draper EA, Wagner DP, Zimmerman JE. APACHE II: A severity of disease classification system. Crit Care Med 1985;13:818e829. 29. Charlson M, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 1987;40:373e383. 30. Aaslid R, Markwalder T, Nornes H. Noninvasive transcranial doppler ultrasound recording of flow velocity in basal cerebral arteries. J Neurosurg 1982;57: 769e774. 31. Krejza J, Mariak Z, Melhem ER, Bert RJ. A guide to the identification of major cerebral arteries with transcranial color Doppler sonography. AJR 2000;174: 1297e1303. 32. McCusker J, Cole M, Abrahamovicz M, et al. Delirium predicts 12-month mortality. Arch Intern Med 2002;162:457e463. 33. Yokota H, Ogawa S, Kurokawa A, Yamamoto Y. Regional cerebral blood flow in delirium patients. Psychiatry Clin Neurosci 2003;57:337e339. 34. Fong TG, Bogardus ST, Daftary A, et al. Cerebral perfusion changes in older delirious patients using 99mTc HMPAO SPECT. J Gerontol A Biol Sci Med Sci 2006;61A:1294e1299. 35. Matta BF, Stow PJ. Sepsis-induced vasoparalysis does not involve the cerebral vasculature: Indirect evidence from autoregulation and carbon dioxide reactivity studies. Br J Anaesth 1996;76:790e794. 36. Terborg C, Schummer W, Albrecht M, et al. Dysfunction of vasomotor reactivity in severe sepsis and septic shock. Intensive Care Med 2001;27: 1231e1234. 37. Thees C, Kaiser M, Scholz M, et al. Cerebral haemodynamics and carbon dioxide reactivity during sepsis syndrome. Crit Care 2007;11:R123.