Accuracy of cerebrospinal fluid Aβ(1-42) for Alzheimer's disease diagnosis: a systematic review and meta-analysis

Unc

orre

cted

Aut

hor P

roof

Journal of Alzheimer’s Disease xx (20xx) x–xxDOI 10.3233/JAD-132264IOS Press

1

Accuracy of Cerebrospinal Fluid A�1-42 forAlzheimer’s Disease Diagnosis: ASystematic Review and Meta-Analysis

1

2

3

Maria Ines Rosaa,∗, Josmar Perucchib, Lidia Rosi Medeirosc, Bruna Fernandesd,Maria Eduarda Fernandes dos Reisb and Bruno Rosa Silvaee

4

5

aLaboratory of Epidemiology at University of Extremo Sul Catarinense, Criciuma, SC, Brazil6

bPostgraduate Program in Health Sciences at University of Extremo Sul Catarinense, Criciuma, SC, Brazil7

cPostgraduate Program in Medicine: Medical Sciences at Federal University of Rio Grande do Sul, Porto Alegre,Brazil

8

9

dGraduate Nutrition at University of Extremo Sul Catarinense, Criciuma, SC, Brazil10

eGraduated in Medicine at University of Extremo Sul Catarinense, Criciuma, SC, Brazil11

Accepted 3 December 2013

Abstract.Background: Alzheimer’s disease (AD) is the most common cause of dementia, accounting for 65–70% of all dementia cases.Objective: We performed a systematic review and meta-analysis to estimate the accuracy of cerebrospinal fluid A�1-42 for thediagnosis of AD. A comprehensive search of the Cochrane Library, MEDLINE, LILACS, Grey literature, and EMBASE wasperformed for papers published from January 1990 to August 2013. The following Medical Subject Headings (MeSH) termswere searched: “Alzheimer disease” or “AD” and “amyloid-�” or “A�1-42”.Methods: We included case-control and cross-sectional studies, prospective or retrospective, that evaluated A�1-42 levels in AD.Statistical analysis was performed using REVMAN 5.2, Meta Disc, and Stata 11.0.Results: A total of 804 citations were identified by the search strategy and 41 studies were included. Meta-analysis showed asensitivity of 84.3% (95% CI: 85.6%–81%) and specificity of 79.4% (95% CI: 77.6%–81.1%). The diagnostic odds ratio was28.9 (95% CI: 21.2–39.5).Conclusion: Our study demonstrated that A�1-42 can discriminate AD from controls with good sensitivity and specificity.

12

13

14

15

16

17

18

19

20

21

22

23

Keywords: A�1-42, Alzheimer’s disease, meta-analyses, systematic review24

INTRODUCTION25

Alzheimer’s disease (AD) is the most common cause26

of dementia, accounting for 65–70% of all dementia27

cases [1]. It is a progressive and insidious dementia28

that severely debilitates and is characterized patholog-29

ically by neuronal loss, extracellular amyloid-� (A�)30

plaques, and intracellular tau neurofibrillary tangles.31

∗Correspondence to: Maria Ines Rosa, Rua Cruz e Souza, 510,Bairro Pio Correa, CEP 88811-550, Criciuma-SC, Brazil. Tel.: +5548 34339976; Fax: +55 48 34335766; E-mail: [email protected].

While the etiology of AD remains controversial, the 32

A� amyloid hypothesis postulates that A� is the pri- 33

mary cause of the disease [2]. AD affects up to 15% of 34

people over the age of 65 years and nearly half of all 35

individuals by the age of 85 years [3]. 36

The diagnosis of AD is based on the identification 37

of dementia with a clinical profile suggestive of the 38

disease from the medical history and clinical examina- 39

tion as well as the exclusion of other causes of dementia 40

using brain imaging and laboratory tests [4]. 41

A�1-42 is considered to be more pathologic as it 42

aggregates more readily and deposits much earlier in 43

ISSN 1387-2877/14/$27.50 © 2014 – IOS Press and the authors. All rights reserved

mailto:[email protected]

Unc

orre

cted

Aut

hor P

roof

2 M.I. Rosa et al. / Alzheimer’s Disease Diagnosis

the disease process than A�1-40 [5]. Low concentra-44

tions of A�1-42 in cerebrospinal fluid (CSF) have also45

been shown to predict the transition of mild cognitive46

impairment to AD and parallel brain A� deposition47

[6, 7].48

The mechanisms controlling brain A� deposition49

and clearance and the relationship between peripheral50

and brain A� concentrations are poorly understood.51

Therefore, we performed a systematic review and52

meta-analysis to evaluate the diagnostic potential of53

CSF levels of amyloid-� protein ending at amino acid54

42 (A�1-42) as biomarkers for AD.55

MATERIALS AND METHODS56

Search strategy57

We searched the Cochrane Library, MEDLINE,58

LILACS, EMBASE, Congress Abstracts, and Grey lit-59

erature (Google Scholar and the British Library) for60

papers published from January 1990 to May 2013. We61

searched using the following terms, both as text words62

and Medical Subject Headings (MeSH) or equivalent63

subject heading/thesaurus terms: “Alzheimer disease”64

or “AD” and “amyloid-�” or “A�1-42”. The search was65

limited to human studies and had no language restric-66

tions. Reference lists of all available primary studies67

were reviewed to identify additional relevant citations.68

The complete search strategy is available on request.69

Screening of abstracts for eligibility70

Abstracts/titles identified from the search were71

screened by three reviewers (MIR, JLP, and LRM).72

Disagreements about study inclusion or exclusion were73

solved by consensus, and when this was not possi-74

ble, they were arbitrarily resolved by a fourth reviewer75

(BRS).76

Study selection77

We included cohort, case-control, and cross-78

sectional studies, prospective or retrospective, that79

evaluated the CSF levels of A�1-42 as a biomarker for80

AD.81

Patients82

We analyzed studies that included patients with83

an initial clinical AD diagnosis according to strictly84

applied clinical diagnostic criteria and in which85

CSF concentrations of A�1-42 had been determined.86

Patients with other causes of cognitive impairment,87

including subdural hematoma, brain tumor, central ner- 88

vous system infection, schizophrenia, major depressive 89

episode, or alcohol abuse, were excluded. 90

Index test 91

The diagnostic test consisted of the analysis of CSF 92

concentrations of A�1-42 in AD. 93

Reference standard 94

The diagnosis of reference was a diagnosis of AD 95

based on the criteria of the Diagnostic and Statistical 96

Manual of Mental Disorders, fourth edition (DSM- 97

IV-TR)1 and the National Institute of Neurological 98

Disorders and Stroke–Alzheimer Disease and Related 99

Disorders (NINCDS–ADRDA) working group [8]. 100

Data abstraction 101

Two investigators (MIR, JLP) independently 102

extracted data from the primary studies. The assess- 103

ment of English-language articles was performed by 104

two reviewers (BRS, BF), while the assessment of 105

non-English articles was independently performed by 106

one reviewer (LRM), following translation when nec- 107

essary. Any disagreement was resolved by consensus 108

for both English and non-English studies. We excluded 109

patients with a diagnosis of mild cognitive impairment. 110

We extracted data on study, patients, and test charac- 111

teristics by using a standardized form. Two reviewers 112

(BRS, JLP) independently abstracted data regarding 113

the prevalence of AD. We calculated the sensitivity, 114

specificity, positive likelihood ratio, and diagnostic 115

odds ratio (DOR) from the primary studies of A�1-42 116

diagnosis. Studies that lacked the data needed to create 117

2 × 2 contingency tables were excluded. Any disagree- 118

ment was resolved by consensus for studies published 119

in all languages. Final inclusion or exclusion was 120

made with reference to a selection criteria check- 121

list. Disagreements about study inclusion or exclusion 122

in meta-analysis were initially solved by consensus, 123

and when this was not possible, they were arbitrarily 124

resolved by one reviewer (LRM). Agreement statistics 125

among reviewers were calculated. 126

Quality assessment 127

All articles meeting the eligibility criteria were 128

assessed for their methodological quality. Quality was 129

assessed with the Diagnostic Accuracy Studies tool 130

recommended by the Cochrane Collaboration, the 131

Quality Assessment of Diagnostic Accuracy Stud- 132

ies (QUADAS-2). This tool comprises four domains: 133

Unc

orre

cted

Aut

hor P

roof

M.I. Rosa et al. / Alzheimer’s Disease Diagnosis 3

patient selection, index test, reference standard, and134

flow and timing. Each domain is assessed in terms of135

risk of bias, and the first three domains are also assessed136

in terms of concerns regarding applicability. Signaling137

questions are included to help judge the risk of bias138

[9]. The quality assessment of the studies was indepen-139

dently performed by MIR and JLP. Any disagreement140

was resolved by consensus.141

Data synthesis and statistical analysis142

For each study, 2 × 2 contingency tables were con-143

structed in which all cases were classified as AD and144

non-AD. We calculated the true-positive rate (TPR;145

sensitivity), specificity, and false-positive rate (FPR;146

1 – specificity) [10]. When 2 × 2 tables had a cell147

with a value of 0, the calculations were corrected, and148

when a study contained two cells with the value of149

0, it was excluded from the analysis [11]. Bivariate150

analysis was used to calculate the pooled estimates151

of sensitivity, specificity, and likelihood ratios (LRs)152

along with 95% confidence intervals (CIs) for the sum-153

mary estimates [12]. The bivariate model preserves the154

2-dimensional nature of the diagnostic data by ana-155

lyzing the logit transformed sensitivity and specificity156

of each study in a single model and considers both157

within-study and between-study variability, in contrast158

to the Littenberg and Moses method, which departs159

from a fixed effects model [13]. To detect the cut-off160

threshold effects, the relationship between sensitiv-161

ity and specificity was evaluated using the Spearman162

correlation coefficient. Pooled estimates were calcu-163

lated only for studies showing sufficient clinical and164

statistical homogeneity. I2 or Q tests (commonly used165

in meta-analysis) are not recommended for assessing166

statistical homogeneity in diagnostic reviews because167

they do not consider the association between sensitiv-168

ity and specificity [14]. The prevalence was calculated169

according to the following equation: TP + FN /number170

total appendicitis extracted from contingency tables.171

The DOR can relate to different combinations of sen-172

sitivities and specificities. The DOR describes the odds173

of the positive test results in participants with disease174

compared with the odds of positive test results in those175

without disease. A single DOR corresponds to a set of176

sensitivities and specificities depicted by the summary177

receiver operating characteristic (SROC) curve. It can178

change according to the threshold and to the ROC curve179

used to define an abnormal examination resulting in an180

expected trade-off between sensitivity and specificity.181

Table 1 shows the formulas and definition commonly182

used test indicators in diagnostic research.183

A SROC curve was generated using data from all 184

thresholds, using the Littenberg and Moses method 185

[13]. Additionally, the area under the curve (AUC) 186

can summarize the inherent capacity of a test for dis- 187

criminating a diseased from a non-diseased subject. 188

Accurate tests usually have AUCs close to 1, and 189

poor tests usually have AUCs close to 0.5 [13]. Sen- 190

sitivity analyses were performed to assess excluded 191

studies with a high risk of verification bias according 192

to QUADAS 2. To analyze publication bias, inverted 193

funnel plots of the logarithmic odds ratio (OR) of indi- 194

vidual studies were plotted against sample size [14]. 195

Statistical analysis was performed with the software 196

Stata 11 (Stata Corp, College Station, TX, USA) soft- 197

ware environment, Meta-DiSc® (Clinical Biostatistics 198

Unit, Ramon y Cajal Hospital, Madrid, Spain) (version 199

1.4), and with RevMan 5.2 [15–17]. 200

RESULTS 201

The search identified a total of 804 citations, of 202

which 419 were potentially relevant after initial eval- 203

uation. From these studies, 378 full articles were 204

excluded. Forty-one (41) primary studies [18–57], 205

involving 5,086 patients (2,932 AD and 2,154 con- 206

trols), met the criteria for inclusion and were analyzed 207

(Table 2). The study selection process is summarized 208

in Fig. 1. 209

Quality assessment 210

The risk of bias for patient selection, index test, 211

reference standard, and flow and timing as well as 212

the concerns for applicability related to the first three 213

domains are shown in Fig. 2. The QUADAS items 214

were evaluated, and 31 studies received a positive 215

assessment in the all domains. Another showed at 216

least one unclear item in the list. Our inter-rater 217

reliability for assessing the methodological quality 218

with QUADAS was 94% (κ = 0.86), indicating good 219

agreement. 220

Sensitivity analysis 221

Among the 41 studies included in the meta-analysis, 222

some were identified as outliers, and one re-analysis 223

was performed without them. However, no significant 224

difference was found in the sensitivity or specificity; 225

thus, we decided not to exclude these papers from the 226

meta-analysis. 227

Unc

orre

cted

Aut

hor P

roof


Table 1Commonly used test indicators in diagnostic research

Reference standardPositive Negative

Positive TP FP TP+FP

Negative FN TNFN+TN

Indextest

TP+FN TN+FP TP+FP+FN+TN

2 x

2 co

ntin

genc

y ta

ble

Test indicator Formula Definition

Sensitivity (true positive rate, TPR) TP/(TP + FN) Proportion negative test results amongthe “healthy”

Specificity (true negative rate, TNR) TN/(TN + FP) Proportion negative test results amongthe “healthy”

Likelihood ratio of a positive test result(LR+)

sensitivity/(1-specificity) Ratio of a positive test result amongdiseased to the same result in the“healthy”

Likelihood ratio of a negative test result(LR−)

(1- sensitivity)/specificity Ratio of a negative test result amongdiseased to the same result in the“healthy”

Diagnostic Odds Ratio (DOR) LR+/LR− = (sensitivity×specificity)/(1–sensitivity) × (1– specificity)= (TP/FN)/(FP/TN)

Summarizes the diagnostic accuracy ofthe index test as a single number thatdescribes how many times higher theodds are of obtaining a test positiveresult in a diseased rather than anon-diseased person.

ROC curve and the area under the curve(AUC)

In a ROC curve the true positive rate(Sensitivity) is plotted in function ofthe false positive rate (100-Specificity)for different cut-off points of aparameter. Each point on the ROCcurve represents asensitivity/specificity paircorresponding to a particular decisionthreshold. The area under the ROCcurve (AUC) is a measure of how wella parameter can distinguish betweentwo diagnostic groups(diseased/normal).

It shows the tradeoff between sensitivityand specificity (any increase insensitivity will be accompanied by adecrease in specificity).

Accuracy is measured by the area underthe ROC curve. An area of 1 representsa perfect test; an area of 0.5 representsa worthless test.

The abbreviations TP, FP, FN, and TN denote the number of true positives, false positives, false negatives, and true negatives, respectively [13].

Fig. 1. Flow chart of trial selection.

Diagnostic performance of CSF concentrations of 228

Aβ1-42 229

Figure 3 shows the forest plot of the DOR and 230

corresponding 95% CI. The DOR showed a signif- 231

icance of 28.9 (95% CI: 21.2–39.5). Meta-analysis 232

showed an overall sensitivity of 84.3% (95% CI: 233

85.6%–81%) and a specificity of 79.4% (95% CI: 234

77.6%–81.1%) (Table 2). Asymmetric SROC curves 235

were used because of the heterogeneity among studies 236

(Fig. 4). The LR+ and LR- were 4.5 (95% CI: 3.7–5.4) 237

and 0.18 (95% CI: 0.14–0.22), respectively. 238

In the subgroup of 11 studies that reported the mean 239

and SD of the A�1-42 levels in the CSF of patients 240

with AD, the mean value was 467.2 pg/ml for AD 241

patients (±189.1), while in controls it was 925.1 pg/ml 242

Unc

orre

cted

Aut

hor P

roof


Table 2Characteristics of the studies included in meta-analysis and data extracted from literature

Study/year Country AD Age Cut off TP FP FN TN Sensibility Specificity(pg/ml)

Andreason et al.,1999 [18]

Sweden 53 – 1130 49 2 4 19 0.92 (0.82–0.98) 0.90 (0.70–0.99)

Andreasen et al.,2001 [19]

Sweden 105 75.9 (±6.8) 523 99 11 6 89 0.94 (0.88–0.98) 0.89 (0.81–0.94)

Andreasen et al.,2003 [20]

Sweden 44 73.7 600 34 4 10 28 0.77 (0.62–0.88) 0.87 (0.71–0.96)

Bibl et al., 2006[21]

Germany 18 69.7 (±10.6) 559 18 1 0 13 1.00 (0.81–1.00) 0.93 (0.66–0.1)

Bibl et al., 2007[22]

Germany 71 555 60 4 11 16 0.84 (0.74–0.92) 0.80 (0.56–0.94)

de Jong et al.,2006 [23]

The Netherlands 61 68 (±8.8) 603 57 2 4 28 0.93 (0.84–0.98) 0.93 (0.78–0.99)

de Meyer et al.,2010 [24]

Belgium 100 76 (±4) 188 91 43 7 71 0.93 (0.86–0.97) 0.62 (0.53–0.71)

Galasko et al.,1998 [25]

USA 82 71.7 (±8.4) 1031 74 12 8 48 0.90 (0.82–0.96) 0.80 (0.68–0.89)

Ganzer et al.,2003 [26]

Germany 105 72 (49–87) 293.5 96 29 9 39 0.91 (0.84–0.96) 0.57 (0.45–0.69)

Gao et al., 2010[27]

Sweden 26 71.84 (±7.3) 150 21 1 5 9 0.81 (0.61–0.93) 0.90 (0.55–0.1)

Hertze et al., 2010[28]

Belgium 94 77 (±7.1) 523 51 4 43 34 0.54 (0.44–0.65) 0.89(0.75–0.97)

Hoglund et al.,2008 [29]

Sweden 22 72 (59–80) – 14 0 8 19 0.64 (0.41–0.83) 1.00 (0.82–1.00)

Ibach et al., 2006[30]

Germany 76 – 530 53 12 23 27 0.70 (0.58–0.80) 0.69 (0.52–0.83)

Ivanoiu & Sindic,2005 [31]

Belgium 75 70.6 (±9.8) – 50 2 25 36 0.68 (0.55–0.77) 0.95 (0.82–0.99)

Kanai et al., 1998[32]

Japan 93 70 (40–92) 256 88 10 5 44 0.95 (0.88–0.98) 0.81 (0.69–0.90)

Kandimalla et al.,2011 [33]

India 44 61.84 (±8.97) 662.65 38 6 4 40 0.90 (0.77–0.97) 0.87 (0.74–0.95)

Kapaki et al.,2001 [34]

Greece 38 68 (±10) 375 33 5 11 36 0.75 (0.60–0.87) 0.88 (0.74–0.96)


Greece 49 67.6 (±9.3) 490 40 10 9 39 0.82 (0.68–0.91) 0.80 (0.66–0.90)


Greece 67 66 (±10) 445 48 8 19 64 0.72 (0.59–0.82) 0.89 (0.79–0.95)


Greece 76 66 (±10) 445 41 10 35 83 0.54 (0.42–0.65) 0.89 (0.81–0.95)

Kester et al., 2010[37]

The Netherlands 47 67.74 (±9) 550 37 6 10 12 0.79 (0.64–0.89) 0.67 (0.41–0. 87)

Koopman et al.,2009 [38]

Belgium 95 76 (71–86) 36 70 19 15 31 0.82 (0.73–0.90) 0.62 (0.47–0.75)

Landau et al.,2010 [39]

EUA 193 78.2 (±7.5) – 159 69 34 160 0.82 (0.76–0.87) 0.70 (0.63–0.76)

Lewcsuk et al.,2004 [40]

Germany 22 68 (62–77) 550 19 6 3 29 0.86 (0.65–0.97) 0.83 (0.66–0.93)

Maddalena et al.,2003 [41]

Switzerland 51 62.3 (±11.2) 490 40 3 11 28 0.78 (0.65–0.89) 0.90 (0.74–0.98)

Morinaga et al.,2010 [42]

Japan 111 71.84 (±7.3) 490 96 7 5 43 0.95 (0.89–0.98) 0.86 (0.73–0.84)

Mulder et al.,2010 [43]

The Netherlands 240 66.7 (±9.2) 550 210 22 38 109 0.85 (0.80–0.89) 0.83 (0.76–0.89)

Noguchi et al.,2005 [44]

Japan 69 70.5 289 58 5 11 38 0.84 (0.73–0.92) 0.88 (0.7 5–0.96)

Olsson et al., 2005[45]

Sweden 78 78 (74–82) 515 71 10 7 43 0.91 (0.82–0.96) 0.81 (0.68–0.91)

Riemenschneideret al., 2002 [46]

Germany 74 68.9 (±9.8) 664 66 2 8 38 0.89 (0.80–0.95) 0.95 (0.83–0.99)

Unc

orre

cted

Aut

hor P

roof


Table 2(Continued)

Study/year Country AD Age Cut off TP FP FN TN Sensibility Specificity(pg/ml)

Roher et al., 2009[47]

USA 47 79.1 (±9.91) – 36 16 9 27 0.8 0 (0.65–0.90) 0.63 (0.47–0.77)

Rosler et al., 2001[48]

Austria 27 68.7 (±2.2) – 46 1 1 20 0.98 (0.89–0.1) 0.95 (0.7–0.1)

Schoonenboom etal., 2004 [49]

The Netherlands 47 59 (52–68) – 46 1 1 20 0.98 (0.89–0.1) 0.95 (0.76–0.1)

Schoonenboom etal., 2005 [50]

The Netherlands 39 – 473 35 2 4 28 0.90 (0.76–0.97) 0.93 (0.78–0.1)

Schimidt et al.,2010 [51]

Germany 76 – 600 64 11 12 45 0.84 (0.74–0.92) 0.80 (0.68–0.90)

Shaw et al., 2009[52]

USA 100 75 (±8) 192 97 27 3 87 0.97 (0.91–0.99) 0.76 (0.67–0.84)

Smach et al., 2008[53]

Tunisie 87 73 505 68 7 9 22 0.88 (0.79–0.94) 0.76 (0.56–0.90)

Tapiola et al.,2000 [54]

Finland 80 71 (±8) 340 63 5 17 34 0.79 (0.68–0.87) 0.87 (0.73–0.96)

Thaweepoksomboonet al., 2011 [55]

Thailand 12 67.79 (12.3) 487 12 8 2 8 0.86 (0.57–0.98) 0.50 (0.25–0.75)

Vanderstichele etal., 2006 [56]

Italy 94 67.5 (60.3–67) 594 87 10 7 50 0.93 (0.85–0.97) 0.83 (0.71–0.92)

Welge et al., 2009[57]

Germany 44 69.8 (±8.49) – 38 31 6 56 0.86 (0.73–0.95) 0.64(0.53–0.74)

TOTAL 2932 2473 444 459 1710 0.84 (0.85–0.81) 0.79 (0.77–0.81)

Patient Selection

Index Test

Reference Standard

Flow and Timing

0% 25% 50%

Risk of Bias Applicability concerns

75% 100% 0% 25% 50% 75% 100%QU

AD

AS

-2 D

om

ain

High

Proportion of studies with low, high, or unclearCONCERNS regarding APPLICABILITY

Proportion of studies with low, high, or unclearCONCERNS regarding APPLICABILITY

Unclear Low

Fig. 2. Results of the evaluation of each study according to QUADAS-2.

(±414.4). There was a statistically significant differ-243

ence between the A�1-42 level in the CSF of patients244

with AD compared to that of the controls (p < 0.001).245

The weighted mean difference was −450.06 pg/ml246

(95% CI: −611.06-289.06) (Fig. 5).247

Assessment of publication bias248

Begg’s funnel plot and Egger’s test were performed249

to assess the publication bias in the literature in all com-250

parison models. The shape of the funnel plot showed251

evidence of obvious asymmetry. The Egger’s test was252

used to provide statistical evidence of funnel plot sym-253

metry (p value for bias- 0.003) (Fig. 6).

DISCUSSION 254

The diagnosis of AD is based on adherence 255

to clinical criteria, such as the National Institute 256

of Neurological and Communicative Disorders and 257

Stroke/Alzheimer’s Disease and Related Disorders 258

Association (NINCDS/ADRDA) and Diagnostic and 259

Statistical Manual of Mental Disorders (DSM-IV) [8]. 260

A more recent set of diagnostic criteria proposed incor- 261

porating imaging in AD diagnosis [4]. Biomarkers 262

are being studied to improve the diagnosis of AD. 263

In this systematic review, we examined the literature 264

regarding CSF A�1-42. In the current study, we found 265

that CSF A�1-42 could discriminate AD from controls 266

Unc

orre

cted

Aut

hor P

roof


Fig. 3. Forest plots of estimates of the diagnostic odds ratio.

with high diagnostic accuracy, showing a sensitivity of267

84.3% (95% CI: 85.6%–81.0%), a specificity of 79.4%268

(95% CI: 77.6%–81.1%), and an OR of 28.9% (95%269

CI: 21.2%–39.5%).270

Several recent studies have demonstrated that low271

CSF A�1-42 is a good indicator of the presence of272

cortical amyloid [58–62]. The pooled sensitivity and273

specificity for A�1-42 in AD versus controls from 13 274

studies involving 600 patients and 450 controls were 275

86% and 90%, respectively [1]. 276

In 2004, a meta-analysis found that the A�1-42/tau 277

ratio had a sensitivity of 71% and specificity of 83% 278

for AD [63]. Rapid cognitive decline was associated 279

with higher CSF-tau and lower A�1-42 levels [64].

Unc

orre

cted

Aut

hor P

roof


Fig. 4. Summary receiver operating characteristic curves.

The major core protein deposited early in senile280

plaques is A�1-42. The CSF of AD patients shows281

decreased levels of A�1-42. There is an approximately282

50% reduction in A�1-42 in AD CSF versus controls283

[65]. The reason for low levels of amyloid protein is284

not clear but it was hypothesized to be the result of285

sequestration of A�1-42 in brain tissue.286

The main neuropathological characteristics of AD287

are the presence of cortical intracellular neurofibril-288

lary tangles and extracellular A� plaques. These cause289

synapse dysfunction, neuronal cell loss, and subse-290

quent brain atrophy [66].291

Fig. 6. Funnel plot of estimated publication bias.

According to the amyloid hypothesis, brain amyloi- 292

dosis, the accumulation of A�1-42 and shorter peptides 293

as well as oligomeric A� assemblies is one of the 294

main causes of neurodegeneration in AD [67]. Each 295

biomarker seems to indicate a specific process in AD. 296

Decreased CSF A�1-42 is an indicator of the brain 297

amyloid burden [68]. 298

AD can be quantified in humans using CSF and PET 299

imaging measurements. A number of recent studies 300

have reported that greater fibrillar A� in cortex, which 301

has been measured previously with amyloid positron 302

emission tomography (PET) imaging using the tracer 303

C-Pittsburgh Compound-B (PiB), is associated with 304

low concentrations of CSF A�1-42 in normal aging 305

and dementia. A recent study describes the temporal 306

relationship between changes in CSF A�1-42 and in 307

PiB-PET signals [69]. The work implies that the lat- 308

Fig. 5. Mean difference between levels of A�1-42 in the cerebrospinal fluid of patients with AD versus controls.

Unc

orre

cted

Aut

hor P

roof


ter changes precede the former. However, conflicting309

findings have also been reported [70, 71] indicating310

that further research is needed to understand how often311

and under what circumstances discordance between312

the two A� markers occurs.313

Compared to other AD biomarkers, PET seems to314

have certain advantages. Compared to magnetic res-315

onance imaging-based markers, PET directly assesses316

the supposed primary molecular event in AD pathogen-317

esis; compared to CSF markers, PET is less invasive318

and allows the determination of biomarker expression319

on a regional level within the brain.320

A limitation of this study was the use of various321

cutoffs among studies; in addition, most studies were322

not separated into early or advanced AD. The con-323

version to AD dementia in mild cognitive impairment324

subjects has shown significant variability among differ-325

ent CSF studies. A longitudinal CSF study in normal326

controls found no change in the A�1-42 levels after327

1 year of follow-up [72]. A few longitudinal studies328

have assessed which CSF biomarker or combination329

of biomarkers improves the prediction of AD dementia330

or cognitive decline.331

In conclusion, our study showed that A�1-42 can332

discriminate AD from controls with a sensitivity of333

84.3% and specificity of 79.4%.334

Two important questions remain: can CSF biomark-335

ers also be used to monitor biological effects of336

treatment and can they predict future AD deterioration337

rate?338

In the future, the integration of more than one imag-339

ing or biomarker modality will increase the accuracy340

of risk prediction and therefore increase the power of341

clinical trials to detect disease modifying or preven-342

tive treatment effects. At the same time, these markers343

will serve to define different stages of the transition344

from healthy aging to AD dementia so that individual345

counseling becomes possible [73].346

ACKNOWLEDGMENTS347

Financial support was provided by the University of348

Extremo Sul Catarinense.349

Authors’ disclosures available online (http://www.350

j-alz.com/disclosures/view.php?id=2045).351

REFERENCES352

[1] Blennow K, de Leon MJ, Zetterberg H (2006) Alzheimer’s353

disease. Lancet 368, 387-403.354

[2] Masters CL, Cappai R, Barnham KJ, Villemagne VL (2006)355

Molecular mechanisms for Alzheimer’s disease: Implications356

for neuroimaging and therapeutics. J Neurochem 97, 1700- 357

1725. 358

[3] Smith MA (1998) Alzheimer disease. Int Rev Neurobiol 42, 359

1-54. 360

[4] Dubois B, Feldman HH, Jacova C, Dekosky ST, Barberger- 361

Gateau P, Cummings J, Delacourte A, Galasko D, Gauthier S, 362

Jicha G, Meguro K, O’brien J, Pasquier F, Robert P, Rossor M, 363

Salloway S, Stern Y, Visser PJ, Scheltens P (2007) Research 364

criteria for the diagnosis of Alzheimer’s disease: Revising the 365

NINCDS-ADRDA criteria. Lancet Neurol 6, 734-746. 366

[5] Iwatsubo T, Odaka A, Suzuki N, Mizusawa H, Nukina H, 367

Ihara Y (1994) Visualization of A�42(43) and A�40 in senile 368

plaques with end-specific A� monoclonals: Evidence that an 369

initially deposited species is A�42(43). Neuron 13, 45-53. 370

[6] Shaw LM, Vanderstichele H, Knapik-Czajka M, Clark CM, 371

Aisen PS, Petersen RC, Blennow K, Soares H, Simon A, 372

Lewczuk P, Dean R, Siemers E, Potter W, Lee VM, Tro- 373

janowski JQ (2009) Cerebrospinal fluid biomarker signature 374

in Alzheimer’s disease neuroimaging initiative subjects. Ann 375

Neurol 65, 403-413. 376

[7] Leow AD, Yanovsky I, Parikshak N, Hua X, Lee S, Toga AW, 377

Jack CR Jr, Bernstein MA, Britson PJ, Gunter JL, Ward CP, 378

Borowski B, Shaw LM, Trojanowski JQ, Fleisher AS, Harvy 379

D, Kornak J, Schuff N, Alexander GE, Weiner MW, Thomp- 380

son PM (2009) Alzheimer’s Disease Neuroimaging Initiative: 381

A one-year follow up study using tensor-based morphome- 382

try correlating degenerative rates, biomarkers and cognition. 383

Neuroimage 45, 645-655. 384

[8] McKhann G, Drachman DA, Folstein M, Katzman R, Price 385

DL, Stadlan EM (1984) Clinical diagnosis of Alzheimer’s 386

disease—report of the NINCDS–ADRDA work group under 387

the auspices of Department of Health and Human Services 388

Task Force on Alzheimer’s disease. Neurology 34, 939-944. 389

[9] Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, 390

Reitsma JB, Leeflang MM, Sterne JA, Bossuyt PM (2011) 391

QUADAS-2: A revised tool for the quality assessment of 392

diagnostic accuracy studies. Ann Intern Med 155, 529-536. 393

[10] Altman DG (1999) Some common problems in medical 394

research. In: Practical Statistics for Medical Research, 9th 395

ed. Altman DG, ed. Chapman, London, pp. 396-439. 396

[11] Irwig L, Tosteson AN, Gatsonis C, Lau J, Colditz G, Chalmers 397

TC, Mosteller F (1994) Guidelines for meta-analyses evalu- 398

ating diagnostic tests. Ann Intern Med 120, 667-676. 399

[12] Reitsma JB, Glas AS, Rutjes AWS, Scholten R, Bossuy PM, 400

Zwinderman AH (2005) Bivariate analysis of sensitivity and 401

specificity produce informative summary measures in diag- 402

nostic reviews. J Clin Epidemiol 58, 982-990. 403

[13] Gatsonis C, Paliwal P (2006) Meta-analysis of diagnostic and 404

screening test accuracy evaluations: Methodologic primer. 405

AJR Am J Roentgenol 187, 271-281. 406

[14] Deeks JJ, Macaskill P, Irwig L (2005) The performance of 407

tests of publication bias and other sample size effects in sys- 408

tematic reviews of diagnostic test accuracy was assessed. J 409

Clin Epidemiol 58, 882e93. 410

[15] Review Manager (RevMan) [Computer program] (2012) Ver- 411

sion 5.2. Copenhagen: The Nordic Cochrane Centre, The 412

Cochrane Collaboration. 413

[16] Zamora J AV, Muriel A, Khan KS, Coomarasamy A (2006) 414

Meta-DiSc: A software for meta-analysis of test accuracy 415

data. BMC Med Res Method 6, 31. 416

[17] Reitsma JB, Rujes AW, Whiting P, Vlassov VV, Leeflang 417

MM, Deeks JJ (2009) Chapter 9: Assessing method- 418

ological quality. In Cochrane Handobook for Systematic 419

Review of Diagnostic Test Accuracy, Version 1.0.0. Deek 420

JJ, Bossuyt PM, Gatsonis C, eds. The Cochrane Col- 421

http://www.j-alz.com/disclosures/view.php?id=2045

Unc

orre

cted

Aut

hor P

roof


laboration, http://srdta.cochrane.org/handbook-dta-reviews,422

Accessed 19 July 2013.423

[18] Andreasen N, Hesse C, Davidsson P, Minthon L, Wallin A,424

Winblad B, Vanderstichele H, Vanmechelen E, Blennow K425

(1999) Cerebrospinal fluid beta-amyloid (1-42) in Alzheimer426

disease: Differences between early- and late-onset Alzheimer427

disease and stability during the course of disease. Arch Neurol428

56, 673-680.429

[19] Andreasen N, Minthon L, Davidsson P, Vanmechelen E,430

Vanderstichele H, Winblad B, Blennow K (2001) Evalua-431

tion of CSF-tau and CSF-Abeta42 as diagnostic markers for432

Alzheimer disease in clinical practice. Arch Neurol 58, 373-433

379.434

[20] Andreasen N, Vanmechelen E, Vanderstichele H, Davidsson435

P, Blennow K (2003) Cerebrospinal fluid levels of total-436

tau, phospho-tau and A beta 42 predicts development of437

Alzheimer’s disease in patients with mild cognitive impair-438

ment. Acta Neurol Scand Suppl 179, 47-51.439

[21] Bibl M, Mollenhauer B, Esselmann H, Lewczuk P,440

Trenkwalder C, Brechlin P, Ruther E, Kornhuber J, Otto M,441

Wiltfang J (2006) CSF diagnosis of Alzheimer’s disease and442

dementia with Lewy bodies. J Neural Transm 113, 1771-1778.443

[22] Bibl M, Mollenhauer B, Lewczuk P, Esselmann H, Wolf S,444

Trenkwalder C, Otto M, Stiens G, Ruther E, Kornhuber J,445

Wiltfang J (2007) Validation of amyloid-beta peptides in CSF446

diagnosis of neurodegenerative dementias. Mol Psychiatry 12,447

671-680.448

[23] de Jong D, Jansen RW, Kremer BP, Verbeek MM (2006)449

Cerebrospinal fluid amyloid beta42/phosphorylated tau ratio450

discriminates between Alzheimer’s disease and vascular451

dementia. J Gerontol A Biol Sci Med Sci 61, 755-758.452

[24] De Meyer G, Shapiro F, Vanderstichele H, Vanmechelen453

E, Engelborghs S, De Deyn PP, Coart E, Hansson O,454

Minthon L, Zetterberg H, Blennow K, Shaw L, Trojanowski455

JQ (2010) Alzheimer’s Disease Neuroimaging Initiative.456

Diagnosis-independent Alzheimer disease biomarker signa-457

ture in cognitively normal elderly people. Arch Neurol 67,458

949-956.459

[25] Galasko D, Chang L, Motter R, Clark CM, Kaye J, Knop-460

man D, Thomas R, Kholodenko D, Schenk D, Lieberburg I,461

Miller B, Green R, Basherad R, Kertiles L, Boss MA, Seu-462

bert P (1998) High cerebrospinal fluid tau and low amyloid463

beta42 levels in the clinical diagnosis of Alzheimer disease464

and relation to apolipoprotein E genotype. Arch Neurol 55,465

937-945.466

[26] Ganzer S, Arlt S, Schoder V, Buhmann C, Mandelkow467

EM, Finckh U, Beisiegel U, Naber D, Muller-Thomsen T468

(2003) CSF-tau, CSF-Abeta1-42, ApoE-genotype and clin-469

ical parameters in the diagnosis of Alzheimer’s disease:470

Combination of CSF-tau and MMSE yields highest sensitivity471

and specificity. J Neural Transm 110, 1149-1160.472

[27] Gao CM, Yam AY, Wang X, Magdangal E, Salisbury C, Peretz473

D, Zuckermann RN, Connolly MD, Hansson O, Minthon L,474

Zetterberg H, Blennow K, Fedynyshyn JP, Allauzen S (2010)475

A�40 oligomers identified as a potential biomarker for the476

diagnosis of Alzheimer’s disease. PLoS One 5, e15725.477

[28] Hertze J, Minthon L, Zetterberg H, Vanmechelen E, Blennow478

K, Hansson O (2010) Evaluation of CSF biomarkers as pre-479

dictors of Alzheimer’s disease: A clinical follow-up study of480

4.7 years. J Alzheimers Dis 21, 1119-1128.481

[29] Hoglund K, Hansson O, Buchhave P, Zetterberg H, Lewczuk482

P, Londos E, Blennow K, Minthon L, Wiltfang J (2008) Pre-483

diction of Alzheimer’s disease using a cerebrospinal fluid484

pattern of C-terminally truncated beta-amyloid peptides. Neu-485

rodegener Dis 5, 268-276.486

[30] Ibach B, Binder H, Dragon M, Poljansky S, Haen E, Schmitz 487

E, Koch H, Putzhammer A, Kluenemann H, Wieland W, 488

Hajak G (2006) Cerebrospinal fluid tau and beta-amyloid 489

in Alzheimer patients, disease controls and an age-matched 490

random sample. Neurobiol Aging 27, 1202-1211. 491

[31] Ivanoiu A, Sindic CJ (2005) Cerebrospinal fluid tau protein 492

and amyloid beta42 in mild cognitive impairment: Prediction 493

of progression to Alzheimer’s disease and correlation with the 494

neuropsychological examination. Neurocase 11, 32-39. 495

[32] Kanai M, Matsubara E, Isoe K, Urakami K, Nakashima K, 496

Arai H, Sasaki H, Abe K, Iwatsubo T, Kosaka T, Watanabe 497

M, Tomidokoro Y, Shizuka M, Mizushima K, Nakamura T, 498

Igeta Y, Ikeda Y, Amari M, Kawarabayashi T, Ishiguro K, 499

Harigaya Y, Wakabayashi K, Okamoto K, Hirai S, Shoji M 500

(1998) Longitudinal study of cerebrospinal fluid levels of tau, 501

A beta1-40, and A beta1-42(43) in Alzheimer’s disease: A 502

study in Japan. Ann Neurol 44, 17-26. 503

[33] Kandimalla RJ, Prabhakar S, Binukumar BK, Wani WY, 504

Gupta N, Sharma DR, Sunkaria A, Grover VK, Bhardwaj N, 505

Jain K, Gill KD (2011) Apo-E�4 allele in conjunction with 506

A�42 and tau in CSF: Biomarker for Alzheimer’s disease. 507

Curr Alzheimer Res 8, 187-196. 508

[34] Kapaki E, Paraskevas GP, Zalonis I, Zournas C (2003) CSF 509

tau protein and beta-amyloid (1-42) in Alzheimer’s disease 510

diagnosis: Discrimination from normal ageing and other 511

dementias in the Greek population. Eur J Neurol 10, 119-128. 512

[35] Kapaki EN, Paraskevas GP, Tzerakis NG, Sfagos C, Seretis A, 513

Kararizou E, Vassilopoulos D (2007) Cerebrospinal fluid tau, 514

phospho-tau181 and beta-amyloid1-42 in idiopathic normal 515

pressure hydrocephalus: A discrimination from Alzheimer’s 516

disease. Eur J Neurol 14, 168-173. 517

[36] Kapaki E, Paraskevas GP, Papageorgiou SG, Bonakis A, 518

Kalfakis N, Zalonis I, Vassilopoulos D (2008) Diagnostic 519

value of CSF biomarker profile in frontotemporal lobar degen- 520

eration. Alzheimer Dis Assoc Disord 22, 47-53. 521

[37] Kester MI, Boelaarts L, Bouwman FH, Vogels RL, Groot ER, 522

van Elk EJ, Blankenstein MA (2010) Diagnostic impact of 523

CSF biomarkers in a local hospital memory clinic. Dement 524

Geriatr Cogn Disord 29, 491-497. 525

[38] Koopman K, Le Bastard N, Martin JJ, Nagels G, De Deyn PP, 526

Engelborghs S (2009) Improved discrimination of autopsy- 527

confirmed Alzheimer’s disease (AD) from non-AD dementias 528

using CSF P-tau (181P). Neurochem Int 55, 214-218. 529

[39] Landau SM, Harvey D, Madison CM, Reiman EM, Foster NL, 530

Aisen PS, Petersen RC, Shaw LM, Trojanowski JQ, Jack CR 531

Jr (2010) Alzheimer’s Disease Neuroimaging Initiative. Com- 532

paring predictors of conversion and decline in mild cognitive 533

impairment. Neurology 20, 230-238. 534

[40] Lewczuk P, Esselmann H, Otto M, Maler JM, Henkel AW, 535

Henkel MK, Eikenberg O, Antz C, Krause WR, Reulbach 536

U, Kornhuber J, Wiltfang J (2004) Neurochemical diagnosis 537

of Alzheimer’s dementia by CSF Abeta42, Abeta42/Abeta40 538

ratio and total tau. Neurobiol Aging 25, 273-281. 539

[41] Maddalena A, Papassotiropoulos A, Muller-Tillmanns B, 540

Jung HH, Hegi T, Nitsch RM, Hock C (2003) Biochem- 541

ical diagnosis of Alzheimer disease by measuring the 542

cerebrospinal fluid ratio of phosphorylated tau protein to beta- 543

amyloid peptide 42. Arch Neurol 60, 1202-1206. 544

[42] Morinaga A, Ono K, Ikeda T, Ikeda Y, Shima K, Noguchi- 545

Shinohara M, Samuraki M, Yanase D, Yoshita M, Iwasa 546

K, Mastunari I, Yamada M (2010) A comparison of the 547

diagnostic sensitivity of MRI, CBF-SPECT, FDG-PET and 548

cerebrospinal fluid biomarkers for detecting Alzheimer’s dis- 549

ease in a memory clinic. Dement Geriatr Cogn Disord 30, 550

285-292. 551

http://srdta.cochrane.org/handbook-dta-reviews

Unc

orre

cted

Aut

hor P

roof


[43] Mulder C, Verwey NA, van der Flier WM, Bouwman FH, Kok552

A, van Elk EJ, Scheltens P, Blankenstein MA (2010) Amyloid-553

beta (1-42), total tau, and phosphorylated tau as cerebrospinal554

fluid biomarkers for the diagnosis of Alzheimer disease. Clin555

Chem 56, 248-253.556

[44] Noguchi M, Yoshita M, Matsumoto Y, Ono K, Iwasa K,557

Yamada M (2005) Decreased beta-amyloid peptide42 in cere-558

brospinal fluid of patients with progressive supranuclear palsy559

and corticobasal degeneration. J Neurol Sci 15, 61-65.560

[45] Olsson A, Vanderstichele H, Andreasen N, De Meyer G,561

Wallin A, Holmberg B, Rosengren L, Vanmechelen E,562

Blennow K (2005) Simultaneous measurement of beta-563

amyloid (1-42), total tau, and phosphorylated tau (Thr181)564

in cerebrospinal fluid by the xMAP technology. Clin Chem565

51, 336-345.566

[46] Riemenschneider M, Wagenpfeil S, Diehl J, Lautenschlager567

N, Theml T, Heldmann B, Drzezga A, Jahn T, Forstl H, Kurz568

A (2002) Tau and Abeta42 protein in CSF of patients with569

frontotemporal degeneration. Neurology 58, 1622-1628.570

[47] Roher AE, Maarouf CL, Sue LI, Hu Y, Wilson J, Beach571

TG (2009) Proteomics-derived cerebrospinal fluid markers572

of autopsy-confirmed Alzheimer’s disease. Biomarkers 14,573

493-501.574

[48] Rosler N, Wichart I, Jellinger KA (2001) Clinical significance575

of neurobiochemical profiles in the lumbar cerebrospinal fluid576

of Alzheimer’s disease patients. J Neural Transm 108, 231-577

246.578

[49] Schoonenboom NS, Pijnenburg YA, Mulder C, Rosso SM,579

Van Elk EJ, Van Kamp GJ, Van Swieten JC, Scheltens P580

(2004) Amyloid beta(1-42) and phosphorylated tau in CSF581

as markers for early-onset Alzheimer disease. Neurology 11,582

1580-1584.583

[50] Schoonenboom NS, Mulder C, Vanderstichele H, Pijnenburg584

YA, Van Kamp GJ, Scheltens P, Mehta PD, Blankenstein MA585

(2005) Differences and similarities between two frequently586

used assays for amyloid beta 42 in cerebrospinal fluid. Clin587

Chem 51, 1057-1060.588

[51] Schmidt D, Zimmermann R, Lewczuk P, Schaller G, Degir-589

menci U, Kreil S, Wiltfang J, Kuwert T, Kornhuber J, Weih590

M (2010) Confirmation rate of blinded (99m)Tc-SPECT591

compared to neurochemical dementia biomarkers in CSF in592

patients with Alzheimer disease. J Neural Transm 117, 1111-593

1114.594

[52] Shaw LM, Vanderstichele H, Knapik-Czajka M, Clark CM,595

Aisen PS, Petersen RC, Blennow K, Soares H, Simon596

A, Lewczuk P, Dean R, Siemers E, Potter W, Lee VM,597

Trojanowski JQ (2009) Alzheimer’s Disease Neuroimag-598

ing Initiative. Cerebrospinal fluid biomarker signature in599

Alzheimer’s disease neuroimaging initiative subjects. Ann600

Neurol 65, 403-413.601

[53] Smach MA, Charfeddine B, Lammouchi T, Dridi H, Ben602

Othman L, Bennamou S, Limem K (2008) Interest of CSF603

beta-amyloid1-42 and t-tau protein level determinations for604

the diagnosis of Alzheimer’s disease. Ann Biol Clin (Paris)605

66, 531-535.606

[54] Tapiola T, Pirttila T, Mehta PD, Alafuzofff I, Lehtovirta M,607

Soininen H (2000) Relationship between apoE genotype and608

CSF beta-amyloid (1-42) and tau in patients with proba-609

ble and definite Alzheimer’s disease. Neurobiol Aging 21,610

735-740.611

[55] Thaweepoksomboon J, Senanarong V, Poungvarin N,612

Chakorn T, Siwasariyanon N, Washirutmangkur L, Udom-613

punthuruk S (2011) Assessment of cerebrospinal fluid (CSF)614

beta-amyloid (1-42), phosphorylated tau (ptau-181) and total615

Tau protein in patients with Alzheimer’s disease (AD) and616

other dementia at Siriraj Hospital, Thailand. J Med Assoc 617

Thai 94, (Suppl 1), S77-S83. 618

[56] Vanderstichele H, De Vreese K, Blennow K, Andreasen N, 619

Sindic C, IvanoiuA, Hampel H, Burger K, Parnetti L, Lanari 620

A, Padovani A, DiLuca M, Blaser M, Olsson AO, Pottel H, 621

Hulstaert F, Vanmechelen E (2006) Analytical performance 622

and clinical utility of the INNOTEST PHOSPHO-TAU181P 623

assay for discrimination between Alzheimer’s disease and 624

dementia with Lewy bodies. Clin Chem Lab Med 44, 1472- 625

1480. 626

[57] Welge V, Fiege O, Lewczuk P, Mollenhauer B, Esselmann H, 627

Klafki HW, Wolf S, Trenkwalder C, Otto M, Kornhuber J, 628

Wiltfang J, Bibl M (2009) Combined CSF tau, p-tau181 and 629

amyloid-beta 38/40/42 for diagnosing Alzheimer’s disease. J 630

Neural Transm 116, 203-212. 631

[58] Fagan AM, Mintun MA, Shah AR, Aldea P, Roe CM, Mach 632

RH, Marcus D, Morris JC, Holtzman DM (2009) Cere- 633

brospinal fluid tau and ptau (181) increase with cortical 634

amyloid deposition in cognitively normal individuals: Impli- 635

cations for future clinical trials of Alzheimer’s disease. EMBO 636

Mol Med 1, 371-380. 637

[59] Tolboom N, van der Flier WM, Yaqub M, Boellaard R, Ver- 638

wey NA, Blankenstein MA, Windhorst AD, Scheltens P, 639

Lammertsma AA, van Berckel BN (2009) Relationship of 640

cerebrospinal fluid markers to 11CPiB and 18F-FDDNP bind- 641

ing. J Nucl Med 50, 1464-1461. 642

[60] Grimmer T, Riemenschneider M, Forstl H, Henriksen G, 643

Klunk WE, Mathis CA, Shiga T, Wester HJ, Kurz A, Drzezga 644

A (2009) Beta amyloid in Alzheimer’s disease: Increased 645

deposition in brain is reflected in reduced concentration in 646

cerebrospinal fluid. Biol Psychiatry 65, 927-934. 647

[61] Jagust WJ, Landau SM, Shaw LM, Trojanowski JQ, Koeppe 648

RA, Reiman EM, Foster NL, Petersen RC, Weiner MW, Price 649

JC, Mathis CA (2009) Relationships between biomarkers in 650

aging and dementia. Neurology 73, 1193-1199. 651

[62] Tapiola T, Alafuzoff I, Herukka SK, Parkkinen L, Hartikainen 652

P, Soininen H, Pirttila T (2009) Cerebrospinal fluid �-amyloid 653

42 and tau proteins as biomarkers of Alzheimer-type patho- 654

logic changes in the brain. Arch Neurol 66, 382-389. 655

[63] Hampel H, Mitchell A, Blennow K, Frank RA, Brettschnei- 656

der S, Weller L, Moller HJ (2004) Core biological marker 657

candidates of Alzheimer’s disease–perspectives for diagnosis, 658

prediction of outcome and reflection of biological activity. J 659

Neural Transm 111, 247-272. 660

[64] Wallin AK, Blennow K, Zetterberg H (2010) CSF biomark- 661

ers predict a more malignant outcome in Alzheimer disease. 662

Neurology 74, 1531-1537. 663

[65] Braak H, Braak E (1991) Neuropathological stage- 664

ing of Alzheimer-related changes. Acta Neuropathol 82, 665

239-259. 666

[66] Ballard C, Gauthier S, Corbett A, Brayne C, Aarsland D, Jones 667

E (2011) Alzheimer’s disease. Lancet 377, 1019-1031. 668

[67] Hardy (2002) the amyloid hypothesis of Alzheimer’s disease: 669

Progress and problems on the road to therapeutics. Science 670

297, 353-356. 671

[68] Riverol M, Lopez OL (2011) Biomarkers in Alzheimer’s dis- 672

ease. Front Neurol 2, 46. 673

[69] Landau SM, Lu M, Joshi AD, Pontecorvo M, Mintun MA, 674

Trojanowski JQ, Shaw LM, Jagust WJ (2013) Alzheimer’s 675

Disease Neuroimaging Initiative. Comparing PET imag- 676

ing and CSF measurements of A�. Ann Neurol, doi: 677

10.1002/ana.23908. 678

[70] Koivunen J, Pirttila T, Kemppainen N, Aalto S, Herukka SK, 679

Jauhianen AM, Hanninen T, Hallikainen M, Nagren K, Rinne 680

JO, Soininen H (2008) PET amyloid ligand [11C] PIB uptake 681

Unc

orre

cted

Aut

hor P

roof


and cerebrospinal fluid beta-amyloid in mild cognitive impair-682

ment. Dement Geriatr Cogn Disord 26, 378-383.683

[71] Tolboom N, van der Flier WM, Yaqub M, Boellaard R,684

Verwey NA, Blankenstein MA, Windhorst AD, Scheltens685

P, Lammertsma AA, van Berckel BN (2009) Relationship686

of cerebrospinal fluid markers to 11C-PiB and 18F-FDDNP687

binding. J Nucl Med 50, 1464.

[72] Vemuri (2010) Serial MRI and CSF biomarkers in normal 688

aging, MCI, and AD. Neurology 75, 143-151. 689

[73] Teipel SJ, Sabri O, Grothe M, Barthel H, Prvulovic D, 690

Buerger K, Bokde AL, Ewers M, Hoffmann W, Hampel H 691

(2013) Perspectives for multimodal neurochemical and imag- 692

ing biomarkers in Alzheimer’s disease. J Alzheimers Dis 33, 693

S329-S347. 694

Accuracy of cerebrospinal fluid Aβ(1-42) for Alzheimer's disease diagnosis: a systematic review and meta-analysis

Documents