Scheltens_edit_Layout 1 27/10/2009 17:29 Page 14
Neurodegenerative Diseases Alzheimer’s Disease
the cerebrospinal fluid (CSF) in AD patients compared with controls time, in each of the three patient groups there was a slight increase
of the same age, with sensitivity and specificity levels ranging from 80 in Aβ
1–42
and tau (4–72 and 49–143pg/ml, respectively), but there was
to 90%. The full reason for the reduced levels in AD is unclear,
7–10
but little change in ptau-181. The differences between patient groups in
it is speculated that Aβ
1–42
aggregates within plaques reducing the these parameters exceeded the changes over time within each group.
freely available fraction detectable in CSF assays. Therefore, the authors concluded that repeatedly monitoring these
biomarkers is not useful because they are insensitive to disease
The main marker of neuronal pathology is the microtubule- progression. Further work showed that change in CSF biomarker
associated tau protein, which is often present in a hyper- levels was not related to change in Mini-Mental State Examination
phosphorylated form in AD patients. Most studies have focused on (MMSE) or to atrophy rate.
23
total tau (T-tau) and hyperphosphorylated tau (P-tau), particularly the
subtypes hyperphosphorylated at threonine 231 (P-tau231) and at One disadvantage of CSF biomarkers is the requirement for a lumbar
threonine 181 (P-tau181). Approximately 50 studies involving 5,000 puncture, a time-consuming, invasive procedure that could restrict
patients have shown that CSF levels of T-tau are increased by the use of CSF biomarkers in large clinical trials and has the side
approximately 300% in AD patients compared with controls, with effect of post-lumbar-puncture headache (PLPH). However, the use
sensitivity and specificity levels similar to those observed for of smaller needles and modern atraumatic techniques has reduced
Aβ
1–42
.
7,10
However, levels of T-tau in control groups typically increase this risk, and the incidence of PLPH has been shown to be lower,
with age, therefore it is a better discriminator in patients <70 years of particularly in elderly subjects with cognitive disturbances.
24
While
age.
11
Increases in P-tau in AD patients have also shown high blood or urine sampling would be more convenient, the much larger
sensitivity and specificity, and are particularly useful in distinguishing volumes of liquid involved mean that T-tau and P-tau levels are
AD from other forms of dementia.
10
diluted and become undetectable by enzyme-linked immunosorbent
assay (ELISA). Aβ peptides can be detected in plasma at
Clinical trials of anti-AD agents require participants with mild concentrations 20–30-fold lower than in CSF, but there is no
symptoms who are likely to display measurable cognitive decline in correlation between Aβ levels in plasma and CSF, and the reduction
the course of a study. Neurochemical biomarkers have the potential of CSF Aβ
1–42
in AD patients is not reflected by a similar reduction in
to identify such individuals and predict AD. Tau protein and Aβ plasma levels.
25
peptides have been used as predictors of AD in those with MCI,
considered a transitional clinical state between normal ageing A recent worldwide multicentre comparison of assays for CSF
and mild AD.
12
Recent studies of CSF in patients exhibiting mild AD biomarkers in AD highlighted another limitation of their use, namely
have found lowered Aβ
1–42
levels, high T-tau or P-tau181 levels or high the high intercentre coefficient of variance (CV). The intercentre
T-tau:Aβ
1–42
ratios, suggesting that these biomarkers may CVs were 30% for Aβ
1–42
, 21% for T-tau and 13% for P-tau in 2004,
quantitatively predict progression of cognitive deficits and dementia.
13
although by 2008 they had fallen to 21, 15 and 9%, respectively.
A previous study found that Aβ
1–42
, but not T-tau or P-tau, may be High intracentre CVs were also observed, at 25, 18 and 7% for
used to predict cognitive decline in healthy elderly individuals.
14
Aβ
1–42
, T-tau and P-tau, respectively. Therefore, there is a high
variation in test results both between and within centres. In order
Combined biomarker levels have been shown to be better diagnostic to reliably demonstrate differences between groups, a CV of <10%
predictors of AD than single measurements. The ratio of Aβ
1–42
to is required. There is a need for standardisation of the analytical
Aβ
1–40
is more accurate than Aβ
1–42
level alone, and the combination procedures employed.
26
of this ratio with T-tau levels further improves accuracy.
15,16
AD has
been distinguished from dementia with Lewy bodies (DLB) by using Several other candidates for AD biomarkers have been investigated
the ratios of Aβ peptides of varying lengths (Aβ
1–42
:Aβ1-38 and but none have as much clinical data as Aβ peptides and Tau.
Aβ
1–42
:Aβ
1–37
) and T-tau.
17
Very high sensitivity and specificity of These include pro-inflammatory cytokines,
27
isoprostanes
28
and β
92 and 89%, respectively, have been reported using combined secretase.
29
A further useful biomarker in AD is serum amyloid P
biomarker measurements.
18
Furthermore, clusters of biomarker levels component (SAP). In a study on 241 patients (67 with AD, 144 with
have been correlated to cognitive profiles of AD patients. A study of MCI and 30 controls), SAP was determined in CSF samples at baseline
177 patients identified a subgroup displaying a distinct cognitive and at follow-up (2.6±1.0 years for AD patients and 2.1±0.8 years for
profile with severe impairment of memory, mental speed and MCI patients).
30
There were no differences in SAP between the AD
executive functions. This was associated with low levels of Aβ
1–42
and MCI groups. Patients with MCI who had developed dementia at
and extremely high levels of T-tau and P-tau181, and showed no link follow-up had lower SAP levels (13mg/l, range 3.3–199.3mg/l) than
to disease duration or severity. These findings have important those not developing it (20.2mg/l, range 7.0–127.7 mg/l; p<0.05). In
implications: AD is a heterogeneous disease and future research is addition, low CSF SAP levels were shown to be correlated with a two-
likely to focus on individualised therapy.
19
fold increased risk of progression to AD (hazard ratio 2.2). It was
suggested that SAP could be used to identify AD patients among a
In order to be clinically meaningful, the intra-individual variation of population with MCI.
biomarker levels over time must be low. A two-year study of 83
patients with MCI analysed levels of T-tau, P-tau181 and Aβ
1–42
and Neuroimaging Techniques
found that intra-individual levels of these biomarkers were highly In vivo neuroimaging techniques have become valuable candidate
stable over this time.
20
In another study including 105 patients biomarkers in the determination of structural changes in the brain
attending a memory clinic, the levels of Aβ
1–42
, tau and ptau-181 in which aid diagnosis of AD. The two most widespread neuroimaging
CSF were measured at baseline and at 21±9 months later (50 patients techniques are magnetic resonance imaging (MRI) and positron
had AD, 38 had MCI and 17 had subjective complaints [SC]).
21,22
Over emission tomography (PET).
14 EUROPEAN NEUROLOGICAL REVIEW