Am j Alzheimers Dis Other Demen-2001-Tsolaki-21-31

7/30/2019 Am j Alzheimers Dis Other Demen-2001-Tsolaki-21-31

1/12

http://aja.sagepub.com/Other Dementias

American Journal of Alzheimer's Disease and

http://aja.sagepub.com/content/16/1/21The online version of this article can be found at:

DOI: 10.1177/153331750101600107

2001 16: 21AM J ALZHEIMERS DIS OTHER DEMENA. Kazis

Tsolaki, V. Sakka, G. Gerasimou, N. Dimacopoulos, O. Chatzizisi, K. N. Fountoulakis, G. Kyriazis, J. Papanastasiotype, other types of dementia, and congrol subjects

relation of rCBF (SPECT), CSF tau, and congnitive function in patients with dementia of the Alzhe

Published by:

http://www.sagepublications.com

can be found at:American Journal of Alzheimer's Disease and Other DementiasAdditional services and information for

http://aja.sagepub.com/cgi/alertsEmail Alerts:

http://aja.sagepub.com/subscriptionsSubscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://aja.sagepub.com/content/16/1/21.refs.htmlCitations:

What is This?

- Jan 1, 2001Version of Record>>

by anca elena on October 11, 2012aja.sagepub.comDownloaded from
http://aja.sagepub.com/http://aja.sagepub.com/http://aja.sagepub.com/http://aja.sagepub.com/content/16/1/21http://aja.sagepub.com/content/16/1/21http://aja.sagepub.com/content/16/1/21http://www.sagepublications.com/http://aja.sagepub.com/cgi/alertshttp://aja.sagepub.com/cgi/alertshttp://aja.sagepub.com/subscriptionshttp://aja.sagepub.com/subscriptionshttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsPermissions.navhttp://aja.sagepub.com/content/16/1/21.refs.htmlhttp://aja.sagepub.com/content/16/1/21.refs.htmlhttp://aja.sagepub.com/content/16/1/21.refs.htmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://aja.sagepub.com/content/16/1/21.full.pdfhttp://aja.sagepub.com/http://aja.sagepub.com/http://aja.sagepub.com/http://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://aja.sagepub.com/content/16/1/21.full.pdfhttp://aja.sagepub.com/content/16/1/21.refs.htmlhttp://www.sagepub.com/journalsPermissions.navhttp://www.sagepub.com/journalsReprints.navhttp://aja.sagepub.com/subscriptionshttp://aja.sagepub.com/cgi/alertshttp://www.sagepublications.com/http://aja.sagepub.com/content/16/1/21http://aja.sagepub.com/


2/12

Abstract

Background. The diagnosis of Alzheimers disease(AD) during life remains difficult and a definite diagno-sis of AD relies on histopathological confirmation at

post-mortem or by cerebral biopsy. It is well known thatlevels of tau proteins are consistently and significantly

increased in the cerebrospinal fluid (CSF) of Alzheimerspatients versus levels in normal controls. However, thesole use of this biochemical marker as a test for AD ishampered by mediocre specificity, since tau concentra-tions may also be elevated in certain other neurologicaldisorders (OND). Studies of the regional cerebral bloodflow (rCBF) are widely performed because of their con-venience and usefulness in a variety of neurological dis-orders. Most studies have reported high diagnosticaccuracy for brain perfusion single-photon emissiontomography (SPECT) in Alzheimers disease.

Methods. In order to improve specificity, in this study,correlation of99mTc-HMPAO SPECT scanning and CSF

tau protein levels was made in 117 patients with AD, 67patients with OND (26 of which had other dementias), and23 age-matched controls. Means and standard deviationsof tau protein levels were 297, 42 221, 12 in AD patientsand 78,07 98, 51 in patients with OND (p = 0.0006). Nocorrelation was noted between CSF tau protein levels andage, duration of the disease, and neuropsychological scoresof mini-mental state examination (MMSE), CambridgeCognitive Examination (CAMCOG), and Functional

21American Journal of Alzheimers Disease and Other DementiasVolume 16, Number 1, January/February 2001

Correlation of rCBF (SPECT), CSF tau,and cognitive function in patients

with dementia of the Alzheimers type,other types of dementia, and control subjects

M. Tsolaki, MD, PhD

V. Sakka, MD

G. Gerasimou, MD

N. Dimacopoulos, MD

O. Chatzizisi, MD

K. N. Fountoulakis, MD, PhDG. Kyriazis, MD

J. Papanastasiou, MD

A. Kazis, MD, PhD

M. Tsolaki, MD, PhD, 3rd Department of Neurology, Aristotle

University of Thessaloniki, Thessaloniki, Greece.

V. Sakka, MD, Department of Neurology, NIMTS, Athens, Greece.

G. Gerasimou, MD, Department of Nuclear Medicine, Aristotle


N. Dimacopoulos, MD, Department of Neurology, NIMTS, Athens,

Greece.

O. Chatzizisi, MD, Department of Immunology, Aristotle University

of Thessaloniki, Thessaloniki, Greece.

K. N. Fountoulakis, MD, PhD, 3rd Department of Psychiatry,

Aristotle University of Thessaloniki, Thessaloniki, Greece.

G. Kyriazis, MD, Department of Immunology, Aristotle University of

Thessaloniki, Thessaloniki, Greece.

J. Papanastasiou, MD, Department of Neurology, NIMTS, Athens,

Greece.

A. Kazis, MD, PhD, 3rd Department of Neurology, Aristot le


http://aja.sagepub.com/http://aja.sagepub.com/http://aja.sagepub.com/http://aja.sagepub.com/


3/12

Rating Scale for Symptoms of Dementia (FRSSD).Findings. There was a bilateral parietal and temporal

hypoperfusion in patients with AD in SPECT in comparisonto normal subjects (p < 0.05) and there was a statisticalcorrelation between this hypoperfusion and neuropsy-chological tests, such as MMSE and CAMCOG (p < 0.01).There was no correlation between tau protein levels andhypoperfusion in SPECT.

Interpretation. Conclusively, the correlation betweenelevated levels of tau proteins and hypoperfusion inSPECT in AD patients therefore cannot improve thespecificity of tests in AD and this means that the determi-nation of CSF tau proteins levels is not a specific diag-nostic test for AD.

Key words: Alzheimers disease, correlation, neu-ropsychological tests, SPECT, tau protein

Introduction

Alzheimers disease (AD) is the leading cause ofdementia in the elderly. AD is the most common of allprogressive degenerative brain diseases leading todementia (about 75 to 80 percent of all cases). The diag-nosis of AD during life, in vivo, remains difficult.Definite diagnosis of AD relies on histopathologicalconfirmation at post-mortem or by cerebral biopsy.

At present, neither the etiology nor the pathogenesisof AD is completely understood. About 50 percent ofcases have a problem in one of four chromosomes: 1, 14,19, and 21.

However, the brains of all patients with AD are char-

acterized by abundant neurofibrillary tangles (NFT),neuropil threads, dystrophic neurites, and senile plaques.NFT represent intracellular accumulations of pairedhelical filaments (PHF), which are also shared with neu-ropil threads and dystrophic neurites. The abundant pres-ence of both senile plaques and tangles in the brains ofAD patients is the only accepted criteria for the unequiv-ocal diagnosis of AD.

Immunocytochemical and biochemical studies showedthat the major component of PHF found in the NFT in thebrain of patients with AD is the microtubule-associatedprotein tau in a highly phosphorylated state.1 These find-ings support the previous observation that an antibody

termed Alz-50 reacted specifically with tangles and aprotein termed A68 on immunoblots of an AD brain.2

Although this protein was found to be clearly elevated inAD brain tissue,3 its presence in cerebrospinal fluid hasbeen difficult to establish. Furthermore, this protein wasshown to be indistinguishable from highly phosphorylatedforms of microtubule-associated tau.4 Additional studieshave shown that tau is elevated in AD brain homogenatesin comparison to control tissues.5 During the last year, two

other studies, one from Sweden6 and the other from Japan,7

found differences in CSF tau protein between controls andAD patients.

On the other hand, measurement of medial temporal lobe

atrophy by computer tomography (CT) or magnetic reso-nance imaging (MRI) has been reported to be a simple andeffective test for AD. These studies were performed inpatients with relatively advanced disease. However,patients in the earliest stages of the disease are believed tobe most likely to benefit from pharmacological treatment.This is the group for whom a reliable diagnostic test couldbe most beneficial. Studies of the regional cerebral bloodflow (rCBF) are widely performed because of their conve-nience and usefulness in a variety of neurological disorders.Most studies have reported a high diagnostic accuracy forbrain perfusion SPECT in Alzheimers disease.8,9

While there is some disagreement among authors,

studies generally show a reduction in temporoparietalperfusion, sometimes accompanied by a reduction infrontal perfusion, with relative sparing of the occipitalcortex and subcortical structures. There is only scant evi-dence of a correlation between measurements of region-al cerebral blood flow and cognitive performance.

The aim of the present study was to establish the potentialutility of the multiple correlation of the pattern of distribu-tion of99mTc-HMPAO with the levels of tau proteins in CSFand neuropsychological tests in the diagnosis of AD.

Material and methods

A total of 207 subjects took part in this study. Theywere patients suffering from dementia of the Alz-heimers type (AD), other types of dementia (OD), neu-rological disorders without dementia (ND), or they werenormal, healthy, elderly control subjects (N). All sub-jects gave written informed consent to participate in thestudy; 93 of them were males (44.9 percent) and 114were females (55.08 percent).

The diagnoses of all subjects are shown on Table 1.

22 American Journal of Alzheimers Disease and Other DementiasVolume 16, Number 1, January/February 2001

Table 1. Composition of the study sample

Diagnosis Count Percent (%)

AD 117 56.52

ND 41 19.80

OD 26 12.56

N 23 11.12

Total 207 100



4/12


Table 2. Age, duration of illness, MMSE, CAMCOG, and the FRSSD scoresas well as rCBF in different areas of brain in AD patients, OD patients, and controls (N)

AD OD N

MeanStandard

deviationMean

Standard

deviationMean

Standard

deviation

Age 67.24 8.31 68.32 9.16

Duration 3.68 1.97 4.71 1.50

MMSE 14.69 6.19 16.65 5.29

CAMCOG 39.70 22.04 50.59 19.81

FRSSD 14.23 8.95 11.42 11.16

Frontal cortex (right) 0.75 0.07 10.66 30.55 0.74 0.04

Frontal cortex (left) 1.69 7.98 0.71 0.08 0.75 0.04

Parietal cortex (right) 0.73 0.09 0.74 0.06 0.78 0.04

Parietal cortex (left) 0.69 0.08 0.69 0.07 0.77 0.04

Temporal (lateral right) 0.69 0.08 0.68 0.12 0.75 0.05

Temporal (middle right) 0.66 0.07 0.66 0.09 0.73 0.06

Temporal (middle left) 0.64 0.09 0.62 0.09 0.73 0.05

Temporal (lateral left) 0.64 0.10 0.62 0.10 0.75 0.05

Occipital cortex (right) 0.90 0.10 0.93 0.05 0.95 0.04

Occipital cortex (left) 0.88 0.10 0.90 0.08 0.95 0.06

Thalamus (right) 0.83 0.08 0.80 0.07 0.85

Thalamus (left)

0.04

0.81 0.07 0.80 0.06 0.85 0.05

Caudate nucleus (right) 0.77 0.06 0.77 0.08 0.79 0.05

Caudate nucleus (left) 0.79 0.07 0.78 0.07 0.79 0.05

Globus pallidus (right) 0.87 0.06 0.87 0.08 0.90 0.04

Globus pallidus (left) 0.85 0.06 0.84 0.07 0.90 0.06



5/12

SPECT study

Study population. Twenty-three subjects were healthycontrol subjects, 117 were AD patients, and 26 were ODpatients. The diagnosis of AD was made in accordancewith National Institute of Neurological and Communi-cative Disorders and Stroke (NINCDS-ADRDA)10 andAmerican Psychiatric Association (DSM-IV))11 criteria.Age, duration of illness, mini-mental state examination(MMSE),12,13 the Cambridge Cognitive Examination(CAMCOG),14,15 and the FRSSD16 scores as well as rCBFin different areas of brain are shown on Table 2. Patientswere either outpatients of the dementia center of G.Papanicolaou Hospital in Thessaloniki or of NIMTSHospital in Athens.

There were no patients with clinically diagnosed de-pression. The Hamilton Depression Rating Scale (DRS)17

score had a mean of 8.33 5.61 and the Geriatric

Depression Scale (GDS)

18

had a mean of 3.74 3.27. Innone of the patients of the AD group was there evidenceof systemic disease and none had a history of head injuryor alcohol and drug abuse.

SPECT data. SPECT data analysis was carried out10 to 45 minutes following an injection of 15 ml (555MBq) 99mTc-HMPAO, using a single-headed rotatingADAC gamma-camera and a dedicated computer. Theinjection of the radio-pharmaceutical was taking place ina quiet room with eyes open and ears unplugged.Acquisition was made by recording 128 angular views(from 0 to 360) , 20 seconds per view, using a 64 64 16 matrix. SPECT data were reconstructed by using a

Butterworth filter with a cut-off frequency of 0.50 cyclescm-1 and corrected for attenuation with a coefficient of0.12. Transverse coronal and sagittal displays were cal-culated from the data set.

A semiquantitative assessment of rCBF was obtainedby creation of rectified Rolandos area (rect Rols) 4 4pixels, representing a brain volume of 16 16 8 mm3.The mean count values of the Rols were measured overthe frontal, parietal, visual, and temporal cortex (medialand lateral) from both hemispheres as well as over thethalami and deep structures of the gray matter (caudatenuclei and putamen-globus pallidus) and lost over thecerebellar hemispheres. The rCBF was calculated as the

relative perfusion toward the gold standard, which wasthe cerebellar hemisphere with the highest activity.

Data analysis. Sensitivity (Sn),19 is the ability of themethod to detect patients, as these are determined by theexternal criterion (clinical diagnosis). Its value is givenfrom the division of true-positive cases (patients correct-ly classified by the instrument, tp) to the total number ofsubjects that were classified as patients by the criterion(true-positive + false-negative).

Specificity (Sp),19 is the ability of the method to detectcontrols, as these are determined by the external criterion(clinical diagnosis). Its value is given from the divisionof true-negative cases (controls correctly classified bythe instrument, tn) to the total number of subjects thatwere classified as controls by the criterion (true-negative+ false-positive).

The Sn and Sp were calculated for the ability of rCBFin the left medial temporal lobe to discriminate betweenAD and N. One-way analysis of variance (ANOVA)20

and discriminant function analysis21 were performed aswell. The Pearson correlation coefficient22 included onlythe AD patients, and was calculated between rCBF andCAMCOG, MMSE, and FRSSD scores.

Study of CSF tau levels

Study population. We studied 117 AD patients, 26 ODpatients, and 41 patients with ND. The diagnosis of ADwas made in accordance to NINCDS-ADRDA andDSM-

IVcriteria. Age and tau levels are shown in Table 3.CSF tau protein. CSF was taken by routine lumbar

puncture after informed consent was obtained from eachpatient or family members. The CSF samples were collect-ed either from the dementia center of G. PapanicolaouHospital in Thessaloniki or NIMTS Hospital in Athens.After CSF was collected , the samples were allocated andstored at -20C until analysis. CSF-tau levels were deter-mined by a previously described sensitive sandwich


Table 3. Age and tau levels

in AD, OD, and ND patients

MeanStandarddeviation

ND patients

Age 47.15 17.83

Tau 92.97 66.40

AD patients

Age 67.24 8.31

Tau 296.49 206.97

OD patients

Age 68.32 9.16

Tau 126.28 100.2



6/12

enzyme-linked immunosorbent assay (Innogenetics,Belgium), according to the manufacturers instructions.

Innotest hTAU-Ag was used for the determination oftau protein in human CSF. Innotest hTAU-Antigen is anenzyme immunoassay (EIA) for the quantitative deter-mination of tau protein.

Data analysis. The Sn and Sp were calculated for theability of CSF tau level to discriminate between AD andOD as well as AD and ND. The Pearson correlation coef-ficient was calculated only for AD patients betweenrCBF and tau level.

Results

SPECT study. The Sn and Sp of the ability of rCBFin the left medial temporal lobe to discriminate betweenAD and N are shown in Table 4. Sn exceeds 0.9 at therCBF level 0.56, and Sp at the level 0.77. So, accordingto these results, we can consider a patient to be suffering

from AD only if rCBF is under 56 percent in the leftmedial temporal lobe, i.e., left hippocampus. Con-versely, we can consider that the subject does not mani-fest AD when rCBF in the same region is over 77percent. An important limitation is that this is true onlywhen the physician must decide whether the patient issuffering from AD or is a healthy subject. Thus, thesecut-off points are not suitable for differentiating betweenAD and other types of dementia.

The mean rCBF values of AD patients, OD patients,and controls (N) are shown in Table 2. The one-wayANOVA revealed that AD and OD differ only in rightfrontal cortex rCBF. This can be due to pure chance, tak-ing into account the number of variables analyzed.However, controls do notdiffer from AD patients only infrontal cortex (bilaterally), right thalamus, caudatenucleus (bilaterally), and left globus pallidus (Table 5).

Discriminant function analysis correctly classified96.15 percent of AD patients and 86.95 percent of con-trols. The classification functions as well as the function(AD function - N function) are shown in Table 6. Whenthe latter function takes values greater than zero, therespective subjects suffer from AD.

The calculation of Pearsons correlation coefficientshowed statistically significant correlation of neuropsy-chological assessment between hypoperfusion in bilater-al parietal lobes and CAMCOG-MMSE (p < 0.001) and

hypoperfusion in left temporal lobe and CAMCOG-MMSE (p < 0.05).

Study of CSF tau levels

The distribution of tau level in the three diagnosticgroups is shown in Figure 1. The Sn and Sp of the abilityof CSF tau level to discriminate between AD and OD aswell as AD and ND are shown in Tables 7 and 8.

A CSF tau value of 660 or greater is highly diagnosticof AD in contrast to other neurological diseases. On thecontrary, a CSF tau value of 175.5 or less is highly infavor of the other neurological diseases. The tau value

156 gives Sn and Sp values around 75 percent and is onlyindicative.

The Pearson correlation coefficient included only ADpatients and was calculated for rCBF and tau level. Nostatistically significant correlation was found. Neitherwas a statistically significant correlation found betweentau proteins levels and MMSE, CAMCOG, or FRSSDscores.

Discussion

The analysis of SPECT data in the current study sug-gests that this examination is capable of differentiating

between AD patients and nondemented control subjects.This is not supported by the international literature, inwhich the use of imaging techniques alone is not suffi-cient to obtain a diagnosis. In our study, these resultsmay be a product of the fact that we included solidlydiagnosed AD patients with a clear decline in cognitivefunction. Also suggested is that all lobes are affectedwith a greater reduction of rCBF in the left hemisphere.This study further suggests that mainly bilateral parietal


Table 4. The Sn and Sp of the abilityof rCBF in the left medial temporal lobe

to discriminate between AD and N

Left medial temporalcortex (rCBF)

Sn Sp

0.53 91.45 0.00

0.56 90.60 0.00

0.57 88.89 0.00

0.58 88.03 0.00

0.59 84.62 0.00

0.75 47.01 65.22

0.76 45.30 82.61

0.77 43.59 91.30



7/12

lobes and secondarily left temporal lobes are correlatedwith neuropsychological instruments, and therefore thereduced rCBF of these structures is mainly responsiblefor cognitive decline. However, the finding of a morewidespread cortical hypoperfusion is in accord with the

presence of a general (noncognitive) psychopathology inAD patients.

The presence of bilateral (symmetrical or asymmetri-cal) posterior parietal cortical defects indicates a highprobability of AD among patients referred for evaluationof dementia.23 According to the authors, patients withnormal perfusion or defects outside the parietal cortexhave a lower probability of AD, but, if the clinical suspi-cion is high and memory and cognition are deteriorating,

these patients should be restudied. Liu and his col-leagues24 have written that the main feature in SPECTstudies in patients with established AD was the bilateral(symmetrical or asymmetrical) reduction of rCBF intemporo-parietal cortex and the presenile onset of thedisease was correlated with unilateral or asymmetricalreduction of rCBF.24 Migneco et al.25 have reported thatthe right temporal cortex was mainly affected in AD.However, some regional CBF disturbances were corre-lated with some concrete functions. Recent study withcombined measurement of regional brain volume andperfusion in AD using registered MRI and SPECT alsosuggests that the most significant changes are found involume (p < 0.001) and perfusion (p < 0.0001) of thetemporo-parietal association cortex.26

Various studies have described that the sensitivity ofSPECT study in early AD is rather low by itself (58 per-cent to 64 percent) and its specificity is about 94 percent.

The sensitivity of SPECT study can be improved byusing the neurologic-psychometric clinical scores.9

Recently, in a study of 114 subjects with histopathologi-cal diagnoses and 105 living controls, SPECT evidenceof parieto-temporal hypoperfusion alone was 89 percentsensitive and 80 percent specific for AD. In the samestudy, CT evidence of medial temporal lobe atrophyalone using published criteria was 89 percent sensitiveand 82 percent specific. When both changes occurred,again in the same case, sensitivity was 86 percent andspecificity was 92.5 percent, yielding an overall accura-cy of 90 percent.27 It is known that sensitivity for clinicaldiagnosis of probable AD has been estimated as 92 per-

cent to 94 percent with positive electron tomography(PET).28 According to our study, the sensitivity andspecificity of SPECT are high enough between AD andcontrols (96.15 percent and 86.95 percent, respectively),while the sensitivity and specificity of SPECT in patientswith AD and OD are low. So, SPECT is a good instru-ment to separate patients with AD from normal controls,but not from patients with other dementias.

In our study there was a strong correlation between ashort neuropsychological scale (MMSE) and perfusiondeterioration of bilateral parietal lobes and left temporallobe (p < 0.01). Also there was a statistical correlationbetween a large neuropsychological scale and perfusion

deterioration of bilateral parietal lobes and left temporallobes (p < 0.01). There are controversies, however, in thisfield. Some authors suggest that addition of rCBF andSPECT can contribute to the diagnosis of blood flow dis-turbances compatible with neuropsychological tests,29-31

and others found no statistical correlation between cog-nitive and perfusion deterioration.32 For better semi-quantitative estimation of relative rCBF, we used rectRols 4 4 pixels normalized to the cerebellum, which is


Table 5. Results of one-way ANOVAbetween AD and OD and AD and N

p

AD/OD

p

AD/N

Frontal cortex (right) 0.009 0.56

Frontal cortex (left) 0.59 0.57

Parietal cortex (right) 0.37 0.011

Parietal cortex (left) 0.91 < 0.001

Temporal (lateral right) 0.65 < 0.001

Temporal (middle right) 0.82 < 0.001

Temporal (middle left) 0.41 < 0.001

Temporal (lateral left) 0.57 < 0.001

Occipital cortex (right) 0.14 0.017

Occipital cortex (left) 0.47 0.001

Thalamus (right) 0.22 0.34

Thalamus (left) 0.62 0.01

Caudate nucleus (right) 0.99 0.20

Caudate nucleus (left) 0.64 0.90

Globus pallidus (right) 0.80 0.09

Globus pallidus (left) 0.63 < 0.001



8/12

never affected in Alzheimers disease or in schizophre-nia;33,34 and in order to improve the sensitivity of ourstudy, we correlated our results with the clinical scores.

With respect to age at onset and SPECT (early andlate onset), in a study with 10 presenile and 16 senilepatients, the perfusion asymmetry provided biologicalevidence for an alteration in left-hemisphere function inpatients for whom the early onset of AD was noted.35

PET studies demonstrating changes in glucose metab-olism and oxygen consumption are very sensitive, butthey are quite expensive and not widely available. Thehigh cost of PET and its limitation to centers withcyclotrons have increased interest in developing theSPECT technique. SPECT imaging has been comparedwith PET for the measurement of CBF.36 SPECT might bea valuable instrument in routine clinical work for differen-tiating AD from normal aging and other dementias.37 The

spatial resolution as well as quantitative accuracy ofSPECT are inferior to those of PET. Recently, tech-netium-99m-hexamethyleno-propylenamine oxime99mTC-HMPAO- Ceretec, Amesham International) has beenused as a potential rCBF agent in the SPECT technique.Our knowledge about the metabolic trapping of theSPECT ligands, such as 99mTC-HMPAO, is not yet com-plete. This is a lipophilic trace, which crosses the blood

brain barrier (BBB) and is almost completely cleanedfrom the blood in a single passage through the compactcerebral circulation. In the brain, this tracer is reformat-ted in an hydrophilic, thus trapped in the brain structuresand its regional distribution is considered proportional torCBF at the time of injection.38

During the past decade, tau has been demonstrated tobe the major protein component of the Alzheimers neu-rofibrillary tangles paired helical filaments (PHF) and


Table 6. Discriminant function analysis correctly classified 96.15 percent of AD patients and 86.95 percentof controls. When AD-N function takes values greater than zero, the respective subject suffers from AD

AD functionp = .55914

N functionp = .24731

AD-N function

Frontal cortex (right) 174.58 153.14 21.44

Frontal cortex (left) 0.61 0.59 0.02

Parietal cortex (right) -131.55 -120.38 -11.17

Parietal cortex (left) 93.74 96.11 -2.37

Temporal (lateral right) 126.03 146.61 -20.59

Temporal (middle right) -164.98 -168.39 3.41

Temporal (middle left) 32.05 10.50 21.55

Temporal (lateral left) 58.31 101.82 -43.51

Occipital cortex (right) 645.34 657.13 -11.79

Occipital cortex (left) -120.24 -128.62 8.37

Thalamus (right) 75.60 66.42 9.18

Thalamus (left) -113.09 -121.00 7.91

Caudate nucleus (right) 29.77 34.36 -4.59

Caudate nucleus (left) -171.51 -186.79 15.27

Globus pallidus (right) -124.43 -131.92 7.49

Globus pallidus (left) 304.01 333.43 -29.42

Constant -321.18 -343.28 22.11



9/12

tau is abnormally hyperphosphorylated in tangles. The

six isoforms of tau are present in a hyperphosphorylatedstate in PHF. It is believed that hyperphosphorylated taucan no longer interact properly with microtubules, lead-ing to cellular disfunction and subsequent neuronaldeath. It is not known whether all tau isoforms serveequally well as substrates for various kinases. Recentresults suggest (1) the presence of N-terminal inserts intau isoforms are important structural determinants thatmodulate the specificity of several kinases ; and (2) the

different tau isoforms may be present at different statesof phosphorylation in PHF.39

Some studies have also shown that tau from differentstages (normal adult brain tau, fetal brain tau, or tau fromAlzheimers brains) contain a number of phosphorylatedresidues, mostly serines and threonines, and in manycases followed by a proline (making them targets of pro-line-directed kinases).

Recent studies suggest that the core component ofAD-associated neurofibrillary tangles (NFT), the micro-tubule-associated protein tau, may be present in CSF.Therefore, tests that reflect the presence of these struc-tures could be helpful in the diagnosis of AD. Cere-brospinal fluid from patients with AD and patients withnon-AD neurological diseases was surveyed by anenzyme immunoassay (EIA) (Innotest hTAU-Ag) toquantify levels of the microtubule-associated protein tauin over 20 studies to date.

Although this protein was found to be clearly elevatedin AD brain tissue, its presence in cerebrospinal fluidhad been difficult to establish until 1993. Tau in CSFseems to have much less heterogeneity than that seenwith the same antibodies against brain homogenates. Atpresent, we do not know if this is due to a preferentialrelease of certain transcriptional or post-translational tauforms.40 By using an extremely sensitive enzyme-linkedimmunosorbent assay (ELISA), low levels of tauimmunoreactivity in CSF were detected for the first timeby a group from Belgium in a number of neurologicalpatients.41 In another study from USA, two years later in1995, monoclonal antibodies specific to tau were made

to confirm the presence of tau in CSF and to define itsutility in the diagnosis of AD. This was done by analyz-ing a well-defined AD patient population and comparingCSF tau levels found in this group with non-AD demen-tia patients as well as healthy and neurological controlpatients. This study strongly suggested that CSF tau lev-els were elevated in patients with probable AD and itwas possible to detect elevated levels of tau in patientsdiagnosed with AD at early clinical stages of the disease(MMSE > 25).42

Two other studies, one from Sweden43 and the otherfrom Japan,44 found differences in CSF tau proteinbetween controls and AD patients. In the first study, the

authors found that the CSF tau concentration was nearlythree times higher in the Swedish apolipoprotein (APP)gene mutation carriers than in healthy noncarriers.Moreover, when these authors compared 21 healthy con-trols with the mutation carriers, they found that the CSFconcentration of tau completely differentiated the groups.In the second study, significantly increased tau levels werefound in AD patients as compared with those in patientswith non-AD neurological diseases and control subjects,


Table 7. Sn and Sp of the ability of various CSFtau levels to discriminate between AD and ND

Tau level Sn Sp

650.000 91.67 41.46

660.000 90.63 41.46

710.000 89.58 41.46

715.000 89.58 43.90

750.000 89.58 48.78

790.000 88.54 48.78

145.300 78.13 68.29

150.300 78.13 70.73

153.000 77.08 70.73

155.500 76.04 73.17

156.000 76.04 75.61

156.300 75.00 78.05

162.600 73.96 80.49

163.000 72.92 82.93

166.800 72.92 85.37

173.000 72.92 87.80

175.500 71.88 90.04

178.600 70.83 92.68



10/12

and increased tau levels were found irrespective of age atonset, apolipoprotein E genotype and clinical stage. Intotal, nine published studies have found that CSF tau lev-els are significantly elevated in AD as compared to eithernondemented controls or patients with other neurologicaldiseases. During the Fifth International Conference onAlzheimers Disease and Related Disorders, many studiesconfirmed these results and suggested combination of tautest with other biochemical markers to improve thespecificity of this AD test.

Several studies using antibodies against the micro-tubule associated protein tau have confirmed its pres-ence in CSF and its elevation in AD patients. Only a fewworks suggest that there is no strong correlation betweenthis rise and clinical-neuropsychological data, such asMMSE45,46 and others, which indicate that tau levels cor-relate with clinical measures of dementia severity.47,48

According to recent information that suggests increased

values and six studies showing decreased values in ADpatients after 15 months of observation, we would ex-pected a good correlation between clinical stages andCSF tau levels.49 In our study, no correlation betweencognitive tests and tau levels was observed.

There are two problems in all of these studies thatdetect tau levels in CSF of patients with dementia. The


Table 8. Sn and Sp of the ability of various CSFtau levels to discriminate between AD and OD

Tau level Sn Sp

66.000 90.63 25.00

70.900 89.58 37.50

130.000 81.25 62.50

138.800 80.21 75.00

139.000 79.17 75.00

302.000 40.63 87.50

302.700 39.58 87.50

308.000 39.58 100.00

Taulevel

Figure 1. Histogram of tau values in the three diagnostic groups.

AD ND OD

Diagnostic group

1400

1200

1000

800

600

400

200

0



11/12

first problem is the overlap of tau concentrations be-tween different groups of demented and nondementedpersons. There are great individual variations in the taulevels within the AD group. The second problem is thatthe diagnosis of AD in all of these studies is only a clini-cal diagnosis and not a post-mortem or confirmation ofthe diagnosis by biopsy; thus, perhaps 10 percent to 20percent of the patients do not have AD.

Conclusions

First, our study supports the finding of bilateral pari-etal lobe and unilateral temporal lobe hypoperfusion inpatients with a clinical diagnosis of AD, in accordancewith NINCDS-ADRDA criteria, and a correlation withcognitive and functional performance rCBF and SPECTfindings adds an objective facet to modern neurologicanalysis and helps in the localization and extension of

the disease. It also suggests that there may be a wide-spread hypoperfusion in AD patients in comparison tonormal subjects and this should be studied, probably inconjunction with general psychopathology.

A serious limitation of the results is that all ADpatients already have been clinically diagnosed; there-fore, no conclusions can be reached relative to the earlydiagnosis of AD at a preclinical stage, which is now nec-essary since new drugs are available.

Second, no statistically significant correlation wasfound between rCBF and tau value. This may mean theCSF tau protein level is a biological marker that is inde-pendent of AD process, SPECT findings, and neuropsy-

chological decline. Since tau does not correlate to eitherneuropsychological assessment or rCBF findings, wecan suggest that those three methods of assessment areindependent of each other. If this is the case, their com-bined use of could improve diagnostic accuracy of AD.

The core of the question, however, is under whichconditions this is true. Whether this is also the case indifferentiating between AD and other types of dementiais unclear and further research is necessary.

References1. Grundke-Iqbal I, Iqbal K, Quilan M, et al.: Microtubule-associatedprotein tau: a component of Alzheimer paired helical filaments.J

Biol Chem. 1986; 261: 6084-6089.2. Wolozin B, Pruchnicki A, Dickson DW, Davies P: A neural antigenin the brains of Alzheimer patients. Science. 1986; 232: 648-650.3. Ghanbari HA, Miller BF, Haigler HJ, et al.: Biochemical assay ofAlzheimers disease-associated protein(s) in human brain tissue.

JAMA. 1990; 263: 2907-2910.4. Ksiezak-Reding H, Davies P, Yen S: Alz-50: A monoclonal anti-body to AD antigens, crossreacts with tau proteins bovine and normalhuman brain.J Biol Chem. 1988; 263: 7943-7946.5. Khatoon S, Grundke-Iqbal I, Iqbal K: Brain levels of microtubule-associated protein tau are elevated in Alzheimers disease: A radio

immuno-blot assay for nanograms of the protein.J Neurochem. 1992;59: 750-753.6. Jensen M, Basun H, Lannfelt L: Increased cerebrospinal fluid tauin patients with Alzheimers disease.Neuroscience Letters. 1995;186: 189-191.7. Hiroyuki A, Terajima M, Miura M, et al.: Tau in cerebrospinalfluid: A potential diagnostic marker in Alzheimers disease. Ann

Neurol. 1995; 38: 649-652.8. Pondreka I, Brucke T, Asenbaum S, et al.: Clinical decision mak-ing and brain SPECT. In Costas DC (ed.):New Trends in Clinical

Neurology and Psychiatry. London: John Libbey and Co., 1993:103-117.9. Goldenberg G, Podreka I, Suess E, Deecke L: The cerebral local-ization of neuropsychological impairment in Alzheimers disease.J

Neurol. 1989; 236: 131-113.10. Report of the NINCDS-ADRDA Work Group: Clinical diagnosisof Alzheimers disease.Neurology. 1984; 34: 939-944.11.Diagnostic and Statistical Manual of Mental Disorders, 4th Ed.(DSM-IV). Washington, DC: American Psychiatric Association,1994:123-163.12. Folstein MF, Folstein SE, McHugh PR: Mini-Mental State: Apractical method for grading the cognitive state of patients for the

clinician.J Psychiatr Res. 1974; 12: 189-198.13. Fountoulakis K, Tsolaki M, Chantzi H, Kazis A: Mini-MentalState Examination (MMSE): A validation study in the Greek elderlypopulation. Encephalos (Greece). 1994; 31: 93-102.14. Roth M, Tym E, Mountjoy CQ, et al.: CAMDEX: A standardisedinstrument for the diagnosis of mental disorder in the elderly withspecial reference to the early detection of dementia.Brit J Psychiat.1986; 149: 698-709.15. Tsolaki M, Fountoulakis K, Chantzi H, Kazis A: The CambridgeCognitive Examination for the elderlyCAMCOG: A validationstudy in the Greek elderly population. Encephalos (Greece), 1996;33: 57-77.16. Hutton TJ: Alzheimers disease. In Cohns Current Therapy.1990: 780.17. Hamilton M: A rating scale for depression.Neurol NeurosurgPsychiat. 1960; 23: 56-62.18. Sheikh JI, Yesavage JA: Geriatric Depression Scale (GDS):Recent evidence and development of a shorter version. ClinGerontol. 1986; 5: 165-171.19. Regier DA, Burke JD: Epidemiology. In Comprehensive Textbookof Psychiatry, V. Baltimore: Williams and Wilkins, 1989; 1: 308-326.20. Altman DG: Practical Statistics for Medical Research. London:Chapman and Hall, 1991: 205-207.21. Altman DG: op. cit.: 324-364.22.Ibid.23. Holman L, Johnson K, Gerada B, et al.: The scintigraphic appear-ance of Alzheimers diseaseA prospective study using 99mTcHMPAO SPECT.J Nucl Med. 1992; 33: 181-185.24. Lie HC, Liu R, Lin KN, et al.: SPECT using 99m Tc HMPAO inAlzheimers disease.Nucl Med Commun. 1992; 13: 535-554.25. Migneco O, Dancourt J, et al.: Major role of right temporal lobe

(RT) involvement in Alzheimers type dementia (ATD). InNuclearMedicine in Research and Practice. Stuttgart: Schattawer, 1992: 442-445.26. Julin P, Jelic V, Shigeta M, et al.: Combined measurement ofregional brain volume and perfusion in Alzheimers disease usingcoregistered MRI/SPECT. Fifth International Conference onAlzheimers Disease and Related Disorders, Osaka, 1996.

Neurobiology of Aging. 1996; 454: S113.27. Jobst KA, Smith AD, Hindley NJ, King E, et al.: SPECT (99m Tc-HMPAO) and X-ray CT in the diagnosis of AD: Improved accuracyover clinical criteria in a cohort of 114 prospectively evaluated subjects




12/12

with histopathological diagnoses. Fifth International Conference onAlzheimers Disease and Related Disorders, Osaka, 1996.

Neurobiology of Aging. 1996; 451: S112.28. Herholz K: FDG PET and differential diagnosis of dementia.

Alzheimer Dis & Assoc Disorders. 1995; 9: 6-16.29. Stergiou V, Grammaticos P: The contribution of nuclear medicineto the diagnosis of Alzheimers disease.Hell Iatr(Greece). 1996;

62:17-24.30. Hanyu H, Abe S, Arai H, et al.: Diagnostic accuracy of singlephoton emission CT in Alzheimer type dementia. Nippon-Ronen-

Igakkai-Zasshi (Japan). 1992; 29(6): 463-468.31. Waldemar G., Walovitch RC, Andersen AR, et al.: 99m Tc-bicisate (neurolite) SPECT brain imaging and cognitive impairmentin dementia of the Alzheimer type: a blinded read of image sets froma multicenter SPECT trial.J Cereb Blood Flow Metabol. 1994; 1(Suppl.14): S99-S105.32. Gomez-Tortosa E, Arias JA, del Barrio A, et al.: SPECT and cog-nitive longitudinal study of patients with dementia of Alzheimer type.Fifth International Conference on Alzheimers Disease and RelatedDisorders, Osaka, 1996.Neurobiology of Aging. 1996; 453:S113.33. Perani D, Di Piero V, et al. Tc 99m HMPAO SPECT study ofregional cerebral perfusion in early Alzheimers disease.J Nucl Med.

1987; 28: 1667-1670.34. Syed GMS, Eagger S, Toone BK, et al.: Quantification of region-al cerebral blood flow (rCBF) using 99m Tc HMPAO SPECT: Choiceof the reference region.Nucl Med Commun. 1992; 13: 811-816.35. Jagust WJ, Reed BR, Seab PJ, Budinger TF: Age at onset and sin-gle-photon emission computed tomographic patterns of regionalcerebral blood flow.Arch Neurol. 1990; 47: 628-633.36. Messa C, Perani D, Lucignani G, et al.: High resolution tech-netium99m-HMPAO SPET in patients with probable Alzheimersdisease: Comparison with fluorine-18-FDG PET.J Nucl Med. 1994;35: 210-216.37. Launes J, Sulkana R, Erkinjutti T, et al.: 99m Tc-HMPAO SPECTin suspected dementia.Nucl Med Commun. 1991; 12: 757-765.38. Ell PJ, Jarritt PH, Costa DC, et al.: Functional imaging of thebrain. Semin Nucl Med. 1981; 3: 214-229.

39. Singh T, Grundke-Iqbal I, Iqbal K: Differential phosphorylationof three-repeat human tau isoforms by several protein kinases. FifthInternational Conference on Alzheimers Disease and RelatedDisorders, Osaka, 1996.Neurobiology of Aging. 1996; 168: S43.40. Vigo-Pelfrey C, Seubert P, Blomquist C, et al.: Tau in CSF: An

antermortem marker for AD? In Iqbal K, Mortimer J, Winblad B,WisnieskyH (eds.)Research Advances in Alzheimers Disease and

Related Disorders. 1995; 22: 197-205.41. Vandermeeren M, Mercken M, Vanmechelen E, et al.: Detectionof tau proteins in normal and Alzheimers disease cerebrospinal fluidwith a sensitive sandwich ELISA. J Neurochem. 1993; 61: 1828-1834.

42. Vigo-Pelfrey C, Seubert P, Blomquist C, et al.: Tau in cere-brospinal fluid: An aftermortem marker for Alzheimers disease?Fourth International Conference on Alzheimers Disease and RelatedDisorders, Minneapolis, 1994. See also Elevation of microtubule-associated protein tau in the cerebrospinal fluid of patients with AD.

Neurology. 1995; 45: 785-793.43. Jensen M, Basun H, Lannfelt L: Increased cerebrospinal fluid tauin patients with Alzheimers disease.Neurosci Lett. 1995; 186: 189-191.44. Hiroyuki A, Terajima M, Miura M, et al.: Tau in cerebrospinalfluid: A potential diagnostic marker in Alzheimers disease. Ann

Neurol. 1995; 38: 649-652.45. Vanmechelen E, Blennow K, Cras P, et al.: Are elevated tau CSFlevels in Alzheimer patients related to clinical parameters? FifthInternational Conference on Alzheimers Disease and Related

Disorders, Osaka, 1996.Neurobiology of Aging. 1996; 674: S168.46. Arai H, Terajima M, Miura M, et al.: Microtubule associated pro-tein tau in CSF: Implications for a potential diagnostic marker in AD.Fifth International Conference on Alzheimers Disease and RelatedDisorders, Osaka, 1996.Neurobiology of Aging. 1996; 666: S166.47. Golombowski S, Muller-Spahn F, Naser W, Hock C: Increasedlevels of tau protein in CSF of patients with AD: Correlation withdegree of cognitive impairment and dependence on ApoE4 allele fre-quency. Fifth International Conference on Alzheimers Disease andRelated Disorders.Neurobiology of Aging. 1996; 665: S165.48. Amberla K, Almkvist O, Basum H, et al.: Tau levels in CSF arerelated to cognitive functions in familial AD. Fifth InternationalConference on Alzheimers Disease and Related Disorders.

Neurobiology of Aging. 1996; 669: 166-167.49. Blomberg M, Jensen M, Basun H, et al.: Increase of CSF tau lev-

els in most Alzheimer patients during 15 months follow up. FifthInternational Conference on Alzheimers Disease and RelatedDisorders.Neurobiology of Aging. 1996; 670: S167.


Am j Alzheimers Dis Other Demen-2001-Tsolaki-21-31

Documents

Transcript of Am j Alzheimers Dis Other Demen-2001-Tsolaki-21-31