ArticlePDF Available

Importance and Impact of Preanalytical Variables on Alzheimer Disease Biomarker Concentrations in Cerebrospinal Fluid

Abstract and Figures

Background: CSF biomarker analyses (β-amyloid (Aβ1-42), total tau (T-tau) and hyperphosphorylated tau (P-tau181P)) are part of the diagnostic criteria of Alzheimer’s disease (AD). Different pre-analytical sample procedures contribute to variability of CSF biomarker concentrations, hampering between-laboratory comparisons. The aim of this study was to explore the influence of fractionated sampling, centrifugation, freezing temperature, freezing delay, and freeze-thaw cycles on CSF biomarker analyses. Methods: Fractionated sampling was studied in sequential aliquots of lumbar CSF. Centrifuged and non-centrifuged samples from the same fraction were compared. CSF samples were subjected to different protocols (liquid nitrogen; -80°C; -20°C; 24h at 2-8°C; 24h and 48h at room temperature). To study the influence of freeze-thaw cycles, samples were thawed up to four times and (re)frozen at -80°C. CSF was collected in polypropylene tubes. CSF biomarker concentrations were determined with commercially available single-analyte INNOTEST assays. Results: CSF biomarker concentrations from non-blood-contaminated samples are not influenced by centrifugation or fractionated sampling. Freezing temperature and delayed storage can affect biomarker concentrations; freezing of CSF samples at -80°C as soon as possible after collection is recommended. Consecutive freezing and thawing of CSF samples for up to three times had little effect. Conclusions: Temperature of freezing, delay until freezing and freeze-thaw cycles significantly influence CSF biomarker concentrations, stressing the need for standard operating procedures for pre-analytical sample handling. The differences observed in this study are, however, relatively small and the impact on the clinical value of these CSF biomarkers needs to be determined.
Content may be subject to copyright.
Importance and Impact of Preanalytical Variables on
Alzheimer Disease Biomarker Concentrations in
Cerebrospinal Fluid
Nathalie Le Bastard,
1,2
Peter Paul De Deyn,
1,3,4
and Sebastiaan Engelborghs
1,3*
BACKGROUND:Analyses of cerebrospinal fluid (CSF) bio-
markers (
-amyloid protein, total tau protein, and hy-
perphosphorylated tau protein) are part of the diagnostic
criteria of Alzheimer disease. Different preanalytical sam-
ple procedures contribute to variability of CSF bio-
marker concentrations, hampering between-laboratory
comparisons. The aim of this study was to explore the
influence of fractionated sampling, centrifugation, freez-
ing temperature, freezing delay, and freeze–thaw cycles
on CSF biomarker analyses.
METHODS:We studied fractionated sampling in sequen-
tial aliquots of lumbar CSF. Centrifuged and noncentri-
fuged samples from the same fraction were compared.
CSF samples were subjected to different protocols (liquid
nitrogen, 80 °C, and 20 °C; 24 h at 2– 8 °C; and 24
and 48 h at room temperature). To study the influence of
freeze–thaw cycles, samples were thawed up to 4 times
and refrozen at 80 °C. CSF was collected in polypro-
pylene tubes. We measured CSF biomarker concentra-
tions with commercially available single-analyte Innotest
assays.
RESULTS:CSF biomarker concentrations from non–
blood-contaminated samples are not influenced by cen-
trifugation or fractionated sampling. Freezing temper-
ature and delayed storage can affect biomarker
concentrations; freezing of CSF samples at 80 °C as
soon as possible after collection is recommended.
Consecutive freezing and thawing of CSF samples up
to 3 times had little effect.
CONCLUSIONS:Temperature of freezing, delay until
freezing, and freeze–thaw cycles significantly influence
CSF biomarker concentrations, stressing the need for
standard operating procedures for preanalytical sample
handling. The differences observed in this study are,
however, relatively small, and the impact on the clinical
value of these CSF biomarkers needs to be determined.
© 2015 American Association for Clinical Chemistry
Cerebrospinal fluid (CSF)
5
biomarkers
-amyloid pro-
tein (A
1–42
), total tau protein (T-tau), and hyperphos-
phorylated tau protein (P-tau
181P
) have been integrated
into the revised diagnostic criteria of Alzheimer disease
(AD) (1). AD CSF biomarker concentrations can show
considerable variation. The sources of this variation can
be found in patient selection and preanalytical, analyti-
cal, or postanalytical aspects of interpretation of the bio-
marker data (2, 3).
Preanalytical confounding factors include the lum-
bar puncture (LP) procedure, tubes for collection and
storage of the sample, and sample handling and storage.
Several studies have been undertaken to quantify their
effects (reviewed by Vanderstichele et al. (2) and del
Campo et al. (4)). Introduction of confounding factors
starts at sample collection. Rostro-caudal gradients for
brain-derived CSF biomarkers are assumed (5). How-
ever, these gradients do not lead to significant differences
in A
1–42
and T-tau concentrations between lumbar
CSF fractions (6), but controversy exists for the compar-
ison of cisternal/ventricular and lumbar CSF (7–9).
Concentrations have not been found to be influenced by
centrifugation (10–12 ), except in 1 recent well-designed
study that demonstrated increased A
1–42
in centrifuged
samples, independent of centrifugation temperature (6).
Studies aimed at unraveling the effect of storage con-
ditions are less clear cut (6, 10, 11, 13–16) owing to dif-
ferent methodological approaches or the lack of statistical
1
Reference Center for Biological Markers of Dementia (BIODEM), Laboratory of Neuro-
chemistry and Behavior, Institute Born-Bunge, University of Antwerp, Antwerp, Belgium;
2
Current affiliation: Fujirebio Europe, Ghent, Belgium;
3
Department of Neurology and
Memory Clinic, Hospital Network Antwerp, Antwerp, Belgium;
4
Department of Neurol-
ogy and Alzheimer Research Center, University Medical Center Groningen, Groningen,
The Netherlands.
* Address correspondence to this author at: UAntwerp, Campus Drie Eiken,
Universiteitsplein 1, 2610 Antwerp, Belgium. Fax +32-3-2652618; e-mail
sebastiaan.engelborghs@uantwerp.be.
Received November 25, 2014; accepted February 18, 2015.
© 2015 American Association for Clinical Chemistry
Previously published online at DOI: 10.1373/clinchem.2014.236679
Previous presentation of the manuscript: Abstract/poster presentation at the 10
th
Interna-
tional Conference on Alzheimer’s and Parkinson’s Diseases (AD/PD 2011), March 9–13,
2011, Barcelona, Spain: Le Bastard N, Luyckx J, Coart E, Vanderstichele H, De Deyn PP,
Engelborghs S. Effect of pre-analytical factors on AD biomarker levels in CSF. Neurodegen
Dis 2011;8(Suppl 1).
5
Nonstandard abbreviations: CSF, cerebrospinal fluid; A
1–42
,
-amyloid protein of 42
aminoacids; T-tau, total tau protein; P-tau
181P
,tau phosphorylated at threonine 181; AD,
Alzheimer disease; LP, lumbar puncture; SOP, standard operating procedure.
Clinical Chemistry 61:5
734–743 (2015) Proteomics and Protein Markers
734
power. Most studies that have evaluated effects of freez-
ing and thawing agree that these manipulations result in
decreased concentrations of A
1–42
; however, there is
disagreement on the number of freeze–thaw cycles ac-
ceptable to maintain A
1–42
concentrations at a constant
level (10–13, 16 –18).
Differences in preanalytical sampling procedures
across studies are summarized in Fig. 1. To facilitate con-
sensus on preanalytical standard operating procedures
(SOPs) for CSF sampling with regard to biomarker anal-
yses in an evidence-based manner, we investigated the
influence of fractionated sampling, centrifugation, freez-
ing temperature, freezing delay, and repeated freeze–
thaw cycles on CSF A
1–42
, T-tau, and P-tau
181P
con-
centrations in 1 experimental setup.
Methods
CSF SAMPLING AND HANDLING
We obtained CSF from 38 patients admitted to the
Memory Clinic of Hospital Network Antwerp between
January and August 2010. All LPs were performed in the
context of a diagnostic workup of presumed cognitive
deterioration. Demographic and clinical data of the pop-
ulation included are described in Supplemental Data,
which accompanies the online version of this article
at http://www.clinchem.org/content/vol61/issue5, and
summarized in online Supplemental Table 1. All patients
and/or their relatives gave informed consent. This study was
approved by local medical ethics committees (University of
Antwerp and Memory Clinic of Hospital Network Antwerp).
LPs were performed in the morning (between 0800
and 1100), with the patient in a sitting position, at the
L3/L4 or L4/L5 interspaces with a 20-gauge, 3.5-inch
Quincke point spinal needle (Becton Dickinson). All pa-
tients were in a fasting state (since the evening before) at
the time of LP. Routine investigation included cell count
and determination of total protein and glucose concen-
trations. The different CSF handling procedures are ex-
plained in detail below and depicted in Fig. 2. We used a
structured study design to evaluate the effect of fraction-
ated sampling (protocol 1: gradient effect), centrifuga-
tion (protocol 2: sample homogeneity), freezing temper-
atures and delayed sample storage (protocol 3), and
freeze–thawing (protocol 4).
Fig. 1. Sampling procedures in different settings (research and clinical routine).
Preanalytical Aspects of CSF Biomarker Analyses for AD
Clinical Chemistry 61:5 (2015) 735
For protocols 1 and 2, CSF from prospectively sam-
pled patients (n 20) was collected into 7 consecutive
polypropylene cryovials (Nalgene® cat. nos. 5000
1020, maximum volume 1.5 mL, and 5000 –0050, max-
imum volume 4.5 mL): 5 fractions (C1–C5) of 1.5 mL
and 2 fractions (C6–C7) of 4.5 mL. The final cryovial
(C7) was gently mixed before pipetting 1.5 mL CSF into
a new cryovial (noncentrifuged C7.1). The same tubes
were used for storage.
To investigate the influence of the fractionated sam-
pling on CSF biomarker concentrations (protocol 1), we
used the C3 and the noncentrifuged C7.1 fractions. The
C3 fraction was chosen over the C1/C2 fractions because
of possible blood contamination in the first milliliter of
CSF. The C3 fraction was also pipetted into another vial
(before freezing) to eliminate possible effects of tube ad-
sorption on only the C7.1 fraction. Fractions C3 and
C7.1 contained CSF taken 7.5 mL apart from each other.
The remaining CSF in the large cryovial C7 was then
centrifuged for 10 min at 1200g(Eppendorf centrifuge
5702, rotor A-4–38), and the supernatant was pipetted
into a new cryovial (C7.1*).
To investigate the influence of centrifugation in
atraumatic CSF (erythrocyte count 500/mm
3
) (proto-
col 2), we measured biomarker concentrations in the
noncentrifuged and centrifuged C7.1 fractions. For pro-
tocols 1 and 2, only CSF samples with erythrocyte
count 500/mm
3
, leukocyte count 10/mm
3
, and
total CSF protein concentration between 12 and 60
mg/dL were included, to analyze samples representa-
tive of healthy CSF. All CSF samples for protocols 1
and 2 were frozen in liquid N
2
at the same time within
2 h (range 20–120 min) after sampling at the hospital
and transported to our Biobank facilities for storage at
80 °C until analysis.
For protocols 3 (n 22) and 4 (n 20), the C6
fraction from prospectively sampled patients with a total
volume of 4.5 mL was centrifuged for 10 min at 1200g
and divided into 10 fractions of 350
L that were trans-
ported at room temperature to our Biobank facilities.
Fig. 2. Study protocol (4 protocols).
In protocol 1, concentrations in the C3 and C7.1* fraction were compared. Because C7.1 was pipetted into a different vial, the C3 fraction was
also pipetted into another vial (before freezing) to eliminate possible effects due to tube adsorption on only the C7.1 fraction. C, cerebrospinal
fluid fraction; RT, room temperature; liq, liquid; f/t, freeze–thaw cycle. The boxes explain the goals of the different protocols.
736 Clinical Chemistry 61:5 (2015)
To investigate the influence of freezing tempera-
ture and freezing delay (protocol 3), 6 CSF aliquots
were frozen in liquid N
2
before storage at 80 °C
(snap freezing), 80 °C (slow freezing, ultralow tem-
perature), or 20 °C (slow freezing, low tempera-
ture); or were incubated at 2–8 °C for 24 2 h, room
temperature for 24 2 h, or room temperature for
48 2 h. Samples incubated at 2–8 °C or room tem-
perature were all transferred to 80 °C after the pro-
posed time period.
To investigate the influence of repeated freeze–thaw
cycles (protocol 4), we used the 4 remaining aliquots of
the C6 fraction. We measured CSF biomarker concen-
trations after a maximum of 4 freeze–thaw cycles. At each
thawing cycle, CSF samples were kept at room tempera-
ture for 2.5 h and refrozen at 80 °C for at least 1 day.
Samples were not vortex-mixed during the thawing pro-
cedure and, to exclude the effect of tube absorption, CSF
samples were also not transferred into another cryovial
after thawing.
CSF BIOMARKER ANALYSIS
We measured CSF concentrations of A
1–42
, T-tau, and
P-tau
181P
with commercially available single-analyte
ELISA kits (Innotest
®
-Amyloid
(1–42)
, Innotest hTau-
Ag, and Innotest Phospho-Tau
(181P)
, Fujirebio Europe)
in April and October 2010. For protocol 3, the mean
delay between sampling and analysis was 47 days (range
14–127). For protocol 4, the mean delay between sam-
pling and analysis was 124 days (range 40–166).
All samples were tested according to the test instruc-
tions provided by the manufacturer. The measurement
ranges of the test kits are described in the package inserts
(A
1–42
, 125–2000 pg/mL; T-tau, 75–1200 pg/mL;
P-tau
181P
, 15.6–500 pg/mL). If the T-tau concentra-
tions obtained were above the highest calibrator concen-
tration of 1200 pg/mL, samples were retested by exten-
sion of the calibrator concentration range through
inclusion of 2400 pg/mL as the highest calibrator con-
centration in the standard curve. Out-of-range bio-
marker values refer to values outside these calibrator
ranges. All samples from 1 patient within 1 protocol were
analyzed during the same ELISA run to exclude test vari-
ability as a potential cause of variation between the dif-
ferent aliquots obtained from 1 patient.
STATISTICAL ANALYSES
CSF biomarker concentrations were log
10
-transformed
before data analysis. Data analysis was performed with
mixed models, controlling for use of different kit lots. We
calculated relative median differences between different
treatments together with the associated 95% CIs. A hy-
pothesis test was considered significant if its associated P
value was 0.050. Analyses were performed with and
without the samples with out-of-range biomarker values.
Only data including the out-of-range samples are pre-
sented in all figures and tables. If the results without the
out-of-range samples were meaningfully different, that is
stated in the text. Statistical analyses were performed with
SAS version 9.2.
Results
FRACTIONATED SAMPLING AND CENTRIFUGATION
(PROTOCOLS 1 AND 2)
Median erythrocyte and leukocyte counts were 1/mm
3
(range 0–245/mm
3
) and 0/mm
3
(range 0 4/mm
3
), re-
spectively. Median protein and glucose concentrations
were 34 mg/dL (range 22–55 mg/dL) and 56 mg/dL
(range 45–79 mg/dL), respectively. The relative median
differences in concentrations of the 3 markers were not
found to be significant for fractionated sampling and
centrifugation (Table 1). The maximal difference be-
tween CSF fractions C3 and C7.1 was approximately 5%
for A
1–42
and P-tau
181P
. For T-tau, uncertainty was
larger (14%) owing to 1 outlier. After excluding this out-
lier, the maximal difference between C3 and C7.1 frac-
tions was approximately 9%. The maximal difference
between CSF fractions C7.1 and C7.1* was approxi-
mately 7% for all 3 biomarkers.
Table 1. CSF biomarker concentrations for protocol 1 (fractionated sampling) and 2 (centrifugation).
Biomarker
Median (range) Protocol 1: C3 and C7.1
(fractions) Protocol 2: C7.1 and C7.1*
(centrifugation)
C3 (n = 20) C7.1 (n = 20) C7.1* (n = 20) Median difference,
% (95% CI) P
a
Median difference,
% (95% CI) P
a
A
1–42
, pg/mL 421 (269–855) 421 (223–907) 426 (246–875) −1.10 (−5.46 to 3.47) 0.615 2.29 (−2.30 to 7.10) 0.315
T-tau, pg/mL 411 (71–2183) 368 (63–1224) 386 (63–1466) 4.32 (−4.91 to 14.45) 0.351 0.91 (−5.01 to 7.19) 0.758
0.53 (−7.33 to 9.06) 0.892
P-tau
181P
, pg/mL 61 (13–239) 59 (13–299) 58 (13–318) −0.18 (−3.94 to 3.72) 0.921 −0.03 (−3.63 to 3.71) 0.988
a
None of the comparisons were statistically significant.
Preanalytical Aspects of CSF Biomarker Analyses for AD
Clinical Chemistry 61:5 (2015) 737
FREEZING TEMPERATURE AND FREEZING DELAY
(PROTOCOL 3)
Median protein and glucose concentrations were 35
mg/dL (range 16–110 mg/dL) and 55 mg/dL (range
41–79 mg/dL), respectively. Samples that needed to be
frozen on the same day (liquid N
2
,80 °C, 20 °C)
were frozen within 4 h after sampling (median 122 min;
range 75–230 min).
Snap freezing of samples in liquid N
2
led to (border-
line) significantly higher A
1–42
concentrations in com-
parison to freezing at 80 °C (P0.048), whereas the
difference between freezing at 20 °C and 80 °C was
nonsignificant (P0.135) (Table 2). Freezing at
20 °C led to significantly lower T-tau and P-tau
181P
concentrations in comparison to freezing at 80 °C
(P0.012 and 0.001, respectively). Excluding the
out-of-range samples, the relative median difference in
T-tau between 20 °C and 80 °C did not remain sig-
nificant (3.37%, 95% CI 0.32% to 6.94%; P
0.073).
CSF A
1–42
concentrations were significantly
higher in samples kept at room temperature for 48 h (P
0.009) and borderline nonsignificant for those kept for
24 h at room temperature (P0.065) in comparison
with samples frozen at 80 °C. CSF P-tau
181P
values
were significantly lower in samples kept at 2–8 °C for
24 h before freezing at 80 °C in comparison with sam-
ples frozen at 80 °C (P0.016). This significant dif-
ference did not hold when excluding the out-of-range
samples (P0.244). There were no significant differ-
ences between freezing at 80 °C and a delay in freezing.
The relative median differences between the different
conditions are visualized in Fig. 3.
FREEZE–THAW CYCLES (PROTOCOL 4)
Median protein and glucose concentrations were 35
mg/dL (range 16–105 mg/dL) and 57 mg/dL (range
41–80 mg/dL), respectively. Samples were frozen within
4 h after sampling (median 115 min; range 75–230 min).
Lower A
1–42
concentrations were found in samples that
underwent 4 freeze–thaw cycles in comparison with all
other samples (1 vs 4, P0.022; 2 vs 4, P0.001; 3 vs
4, P0.002) (Fig. 4). The T-tau concentrations were
also lower after 4 freeze–thaw cycles in comparison with
samples that underwent 2 freeze–thaw cycles (P
0.016), but were not significantly different from samples
with 1 or 3 freeze–thaw cycles. Freezing and thawing of
CSF had no effect on P-tau
181P
concentrations.
Discussion
Poor standardization of preanalytical sample handling
procedures has hampered the comparison of CSF
A
1–42
, T-tau, and P-tau
181P
concentrations between
different laboratories or studies. In this study, we exam-
ined the effects of fractionated sampling, centrifugation,
freezing temperature, freezing delay, and freeze–thaw cy-
cles. CSF biomarker concentrations from non–blood-
contaminated samples were not found to be significantly
influenced by centrifugation or fractionated sampling.
Table 2. CSF biomarker concentrations for protocol 3 (freezing temperature and delay) and 4 (freeze–thaw cycles).
a
Protocol A
1–42
, pg/mL
Pvs
standard
procedure T-tau, pg/mL
Pvs
standard
procedure
P-tau
181P
,
pg/mL
Pvs
standard
procedure
Protocol 3: freezing temperature
and delay (n = 22)
−80°C
b
300 (165–579) 359 (46–1875) 53 (14–172)
−20°C 313 (166–636) 0.135 348 (52–1550) 0.012 52 (10–165) 0.005
Liquid N
2
335 (172–594) 0.048 396 (63–1604) 0.405 55 (10–173) 0.068
24 h at 2–8°C 307 (145–576) 0.624 386 (52–1890) 0.296 50 (7–169) 0.016
24 h at room temperature 324 (178–612) 0.065 403 (48–1871) 0.732 55 (10–171) 0.625
48 h at room temperature 320 (195–558) 0.009 384 (50–1919) 0.804 54 (8–169) 0.180
Protocol 4: freeze–thaw cycles
(n = 20)
1 freeze–thaw cycle
b
428 (190–837) 417 (89–2063) 48 (14–162)
2 freeze-thaw cycles 443 (229–892) 0.667 422 (90–2228) 0.133 47 (14–155) 0.742
3 freeze-thaw cycles 411 (229–939) 0.783 413 (86–1928) 0.672 49 (14–171) 0.245
4 freeze-thaw cycles 398 (178–833) 0.022 406 (80–2070) 0.491 50 (15–162) 0.749
a
Data are medians (range). Significant differences are indicated in bold. See also Figure 2.
b
Standard procedure.
738 Clinical Chemistry 61:5 (2015)
Different freezing temperatures and delays in the freezing
process introduced differences compared with freezing at
80 °C, which is considered to be the standard proce-
dure. The consecutive freezing and thawing of CSF sam-
ples for up to 3 times demonstrated little effect on bio-
marker concentrations.
FRACTIONATED SAMPLING
The volume of CSF withdrawn varies across centers, pa-
tients (evacuating LP in case of normal-pressure hydro-
cephalus), or study type (research, routine, clinical trial),
and CSF can be collected in consecutive polypropylene
vials or in a large polypropylene tube before aliquoting.
Fig. 3. Relative median differences (%) between the differ-
ent freezing conditions and introductions of a delay before
freezing, compared with freezing at −80 °C.
Results are presented as the relative median difference with
95% CI. For example, the relative median difference for
A
1–42
comparing samples that remained at room tempera-
ture (RT) for 48 h and samples that were frozen immediately at
−80 °C is approximately 8%, indicating that the A
1–42
con-
centrations are 8% higher in samples that were kept at room
temperature for 48 h.
Fig. 4. Relative median differences (%) between the differ-
ent freeze–thaw cycles.
Results are presented as the relative median difference with 95%
CI. For example, the relative median difference for A
1–42
com-
paring samples that underwent 1 and 4 freeze–thaw cycles is
approximately 8%, indicating that the A
1–42
concentrations
were 8% lower in the samples that underwent 4 freeze–thaw cy-
cles in comparison to samples that underwent only 1 freeze–thaw
cycle. 1/2, 1 vs 2; 1/3, 1 vs 3; 1/4, 1 vs 4; 2/3, 2 vs 3; 2/4, 2 vs 4;
3/4,3vs4.
Preanalytical Aspects of CSF Biomarker Analyses for AD
Clinical Chemistry 61:5 (2015) 739
Brain-derived proteins usually show a rostro-caudal con-
centration gradient, with higher concentrations in ven-
tricular CSF compared with lumbar CSF (5, 8), imply-
ing that different volumes or fractionated sampling could
generate differences in CSF biomarker concentrations.
However, differences in A
1–42
and T-tau between lum-
bar CSF fractions were not found (6). Our data on lum-
bar CSF fractions agree with previous studies and showed
that this is also true for P-tau
181P
, although a difference of
only 7.5 mL CSF was investigated (between 2 fractions),
and the estimated lumbosacral CSF volume of healthy
individuals, though highly variable, is much more than
7.5 mL (19).
CENTRIFUGATION
In the case of blood-contaminated CSF, freezing of sam-
ples without prior centrifugation might lead to hemoly-
sis, which is known to cause aberrant results in the
determination of several other substances present in
erythrocytes (e.g., neuron-specific enolase (20) or sy-
nuclein (21)). Blood-contaminated CSF was defined as
500 erythrocytes/mm
3
, and the detection limit of vi-
sual inspection of CSF for blood contamination is ap-
proximately 0.05% (vol/vol) blood (22). As a conse-
quence of blood–brain barrier deficits, abundant plasma
proteins could influence the outcome of CSF tests, al-
though it was shown in the past that the ex vivo addition
of a number of plasma proteins followed by a direct mea-
surement in the assay or overnight incubation of the sam-
ple did not affect A
1–42
concentrations, except for con-
jugated bilirubin and fibrinogen (17).
Therefore, we wondered whether centrifugation of
CSF is mandatory for non–blood-contaminated samples.
Our results indicate that centrifugation has no effect on
CSF biomarker concentrations in macroscopically non–
blood-contaminated samples. For A
1–42
, 2 studies that
did not freeze–thaw their CSF samples before analysis
yielded contradictory results; one showed increasing
A
1–42
concentrations in centrifuged samples indepen-
dent of the centrifugation temperature (4 °C or room
temperature) (6), whereas the other showed stable
A
1–42
concentrations independent of the time lapse (1,
4, 48, or 72 h) between collection and centrifugation
(11). Another study on A
1–42
that did freeze the cen-
trifuged and noncentrifuged samples before analysis
showed no changes in A
1–42
, independent of the tech-
nology platform (12). All studies investigating the effect
of centrifugation on T-tau and P-tau
181P
showed no dif-
ferences (10–12 ). Moreover, no significant changes in
A
1–42
arose from the addition of 5000 lysed erythro-
cytes/mm
3
(6). The concentrations of A
1–42
and T-tau
were also determined in 1 severely blood-contaminated
sample (28.800 erythrocytes/mm
3
) that had been di-
vided, with 1 aliquot being centrifuged before freezing
and 1 aliquot not centrifuged, which did not lead to a
considerable difference (11). The present data, com-
bined with the majority of the data from the literature,
allows us to conclude that centrifugation (before or after
freezing) does not affect CSF biomarker concentrations,
probably not even when CSF is contaminated with
blood. Although A
1–42
can also be found in plasma,
originating from different pools such as platelets, muscle,
and liver, its concentration is much lower than in CSF
(23) and is therefore unlikely to significantly alter the
concentrations measured in blood-contaminated CSF.
Maximal differences of 9% and 7% for fractionated sam-
pling and centrifugation, respectively, were smaller than
the reported interassay CV for the assays used (10, 17).
In addition, centrifugation of the non–blood-
contaminated samples did not lead to a significantly dif-
ferent intraassay CV for A
1–42
(4.6% vs 3.8%), T-tau
(4.1% vs 4.8%), or P-tau
181P
(2.2% vs 1.4%). However,
the possible impact of hemoglobin concentrations and/or
the number of red blood cells on assay performance has
not been studied in detail.
FREEZING TEMPERATURE
Freezing of samples can be done at different tempera-
tures: 20 °C, 80 °C, and 196 °C (liquid N
2
) are
most commonly used. From long-term stability studies,
it is already known that A
1–42
concentrations in sam-
ples frozen at 80 °C are stable for at least 2 years
(6, 18), as are concentrations of T-tau and P-tau
181P
when stored at 20 °C for at least for 2 and 4 years,
respectively (12, 18). A storage artifact is seen for cystatin
C analysis after freezing at 20 °C, but not at 80 °C
(24). Long-term stability of CSF samples in liquid N
2
,or
direct freezing of a sample in liquid N
2
compared with
other freezing conditions, has never been investigated
extensively. Freezing at 80 °C in comparison to (ini-
tial) freezing in liquid N
2
resulted in borderline signifi-
cantly lower A
1–42
concentrations, whereas the com-
parison of 80 °C with 20 °C was not found to be
significant. We can speculate that freezing in liquid N
2
quickly reduces degradation of proteins by protease ac-
tivity and stabilizes all proteins in the CSF, thereby pre-
venting loss of full-length A
1–42
. However, freezing in
liquid N
2
immediately after CSF collection should then
have led to higher A
1–42
concentrations in comparison
to freezing in liquid N
2
after 2 h, which was shown not to
be the case (14), and freezing at 80 °C should have also
resulted in higher concentrations than freezing at
20 °C. Indeed, T-tau and P-tau
181P
concentrations
differed significantly between 80 °C and 20 °C fro-
zen samples, but not those of A
1–42
. Although a signif-
icant difference was found for A
1–42
after freezing in
liquid N
2
in comparison to freezing at 80 °C, these
results should be interpreted with caution, because of the
relatively large CI for A
1–42
in comparison to T-tau and
P-tau
181P
, indicative of a larger variation during measure-
740 Clinical Chemistry 61:5 (2015)
ment for A
1–42
. One other study evaluating the effect of
different freezing temperatures for A
1–42
found no dif-
ference (6). In conclusion, freezing in liquid N
2
yields
the highest concentrations for all 3 markers but is not the
most practical solution, and freezing of CSF samples at
20 °C pending biomarker analysis, which would be
preferred from a practical point of view, is discouraged.
The fact that long-term studies show stable concentra-
tions for all 3 biomarkers means that differences between
liquid N
2
,80 °C, and 20 °C are most likely attribut-
able to the initial freezing conditions. This is also seen for
tubes that immediately adsorb A
1–42
(25).
FREEZING DELAY
CSF samples are frequently shipped to reference labora-
tories at room temperature, with cooling, or after freezing
(Fig. 1). Shipment of samples by means of regular mail
takes at least 24 h. Therefore, A
1–42
, T-tau, and
P-tau
181P
should preferably be stable at room tempera-
ture for several days (also because shipment of CSF sam-
ples on dry ice is very costly). In comparison with imme-
diate (within 4 h) freezing of samples at 80 °C, the
A
1–42
concentration gradually increased by almost 15%
during 48 h of storage at room temperature. Bjerke et al.
(6 ) stated that storage at room temperature for 24 h did
not significantly affect A
1–42
concentrations in compar-
ison to both fresh samples and storage at 80 °C until
analysis. Another study found relatively stable CSF bio-
marker concentrations even after 4 days, although unsys-
tematic variation increased over time (18). In contrast to
this, A
1–42
has been found to decrease during the first
48h(11) or, more in line with the present study, even to
increase within 24 h (14). An increase could be explained
by deoligomerization of A
1–42
or the release of A
1–42
from amyloid-binding proteins, such as (pre)albumin or
(apo)lipoproteins. The largest proportion of A
in
plasma and CSF is bound to proteins (26, 27). A higher
ratio of free to bound A
could substantially alter A
1–42
concentrations. Incubation of the assay at different tem-
peratures influences the outcome of the A
1–42
assay,
with higher A
1–42
concentrations at higher tempera-
tures (17), and storage of samples at higher temperatures
does seem to have the same effect on A
1–42
(24hat
2–8 °C or room temperature, P0.020), in contrast to
what has previously been published on a smaller popula-
tion (11). In addition, Sancesario et al. (15 ) also demon-
strated (reversible) increased A
1–42
in samples that were
kept at 37 °C before freezing, an effect that was seen for
samples from AD patients, not controls. T-tau was found
to be stable at room temperature (14, 16, 28), for up to
22 days in 1 particular study, but tended to decrease in
that same study after 12 days when stored at 37 °C de-
grees (11), which would indicate that stability of T-tau is
also dependent on temperature. No effects of storage at
room temperature on P-tau
181P
were found in previous
studies (10, 14, 16). One study that examined possible
differences between storage at 4 °C for 4, 24, and 72 h
and immediate freezing at 80 °C could not demon-
strate an effect (10). In conclusion, we found significant
differences in A
1–42
(and P-tau
181P
) concentrations af-
ter delayed storage before freezing.
FREEZE–THAW CYCLES
Our results showed a maximal decrease of 16% in A
1–42
concentration after the fourth freeze–thaw cycle, which is
consistently different from all other freeze–thaw cycles.
Previous studies also found decreasing A
1–42
after the
third (20%) or fifth (15%) freeze–thaw cycle
(11, 12, 17), whereas no differences were found between
fresh CSF samples and samples frozen and thawed once
(11, 13, 17). T-tau showed a (nonsignificant) increase
after 2 freeze–thaw cycles, after which a decreasing trend
was imposed, resulting in a significant difference between
freeze–thaw cycles 2 and 4 (7%). T-tau is considered to
be very stable, as several authors failed to find significant
differences in T-tau concentrations between samples fro-
zen and thawed once (11, 13 ) and in up to 6 freeze–thaw
cycles (11, 12, 16), although Schoonenboom et al. (11 )
also observed an increase in T-tau followed by a decrease.
P-tau
181P
, on the other hand, proved to be very stable in
the present study, and the small CIs point to a low vari-
ability in evolution of P-tau
181P
among the different in-
dividuals, in accord with the results of earlier studies
(10, 16, 29). Zimmermann et al. (18 ) found no signifi-
cant systematic changes of the biomarker concentrations,
but did observe increasing unsystematic variation after 3
freeze–thaw cycles, especially for A
1–42
and T-tau. In
summary, A
1–42
and possibly also T-tau are influenced
by repeated freeze–thaw cycles.
LIMITATIONS OF THE STUDY
Physiological variability was limited through fractionated
sampling in a relatively small number of patients, which
is both a strength (variability of the results is largely due
to variability of the preanalytical variables under study)
and a limitation (the limited physiological variability
might not necessarily reflect daily clinical practice).
Moreover, the impact of the preanalytical variables might
depend on the brain pathology (AD vs non-AD demen-
tias vs controls). Therefore, the lack of a control group is
another limitation of the study. A replication study in a
larger and more heterogeneous population including a
control group might strengthen the conclusions of this
article.
Conclusions
The following recommendations can be proposed as a
result of the present study.
Preanalytical Aspects of CSF Biomarker Analyses for AD
Clinical Chemistry 61:5 (2015) 741
Fractionated sampling: It is possible to collect CSF in 1
large volume (for division into aliquots at a later phase)
or several smaller volumes, since the total volume of
CSF collected does not affect the concentrations of tau
and amyloid proteins.
Centrifugation: If atraumatic samples are used, centrif-
ugation is not required for CSF biomarker analyses.
Freezing temperature: Freezing at 80 °C can be a
general recommendation for long-term as well as
short-term storage, since freezing at 20 °C seems to
influence CSF biomarker concentrations. Freezing in
liquid N
2
is not recommended.
Freezing delay: The delay in freezing should be
minimized.
Freeze–thaw cycles: One freeze–thaw cycle just before
analysis is standard procedure, but 1 or 2 additional
freeze–thaw cycles are allowed.
Author Contributions: All authors confirmed they have contributed to
the intellectual content of this paper and have met the following 3 require-
ments: (a) significant contributions to the conception and design, acquisi-
tion of data, or analysis and interpretation of data; (b) drafting or revising
the article for intellectual content; and (c) final approval of the published
article.
Authors’ Disclosures or Potential Conflicts of Interest: Upon man-
uscript submission, all authors completed the author disclosure form. Dis-
closures and/or potential conflicts of interest:
Employment or Leadership: N. Le Bastard, Fujirebio Europe.
Consultant or Advisory Role: None declared.
Stock Ownership: None declared.
Honoraria: None declared.
Research Funding: This work is part of the BIOMARKAPD project
within the EU Joint Programme for Neurodegenerative Disease Re-
search (JPND). The work was supported by the Research Foundation
Flanders (FWO), the Agency for Innovation by Science and Technol-
ogy (IWT), the Belgian Science Policy Office Interuniversity Attraction
Poles (IAP) program, and the Flemish government–initiated Methusa-
lem excellence grant. Innotest assays were provided by Fujirebio Europe
(Ghent, Belgium). N. Le Bastard, Alzheimer’s Biomarkers Standardiza-
tion Initiative (ABSI); P.P. De Deyn, the central Biobank facility of the
Institute Born-Bunge/University Antwerp; S. Engelborghs, the Univer-
sity of Antwerp Research Fund, the Alzheimer Research Foundation
(SAO-FRA), ABSI, and the EU/EFPIA Innovative Medicines Initiative
Joint Undertaking (EMI grant no. 115372).
Expert Testimony: None declared.
Patents: None declared.
Role of Sponsor: The funding organizations played no role in the
design of study, choice of enrolled patients, review and interpretation of
data, or preparation or approval of manuscript.
Acknowledgments: The authors thank Ellis Niemantsverdriet, MSc,
Charisse Somers, MSc (BIODEM, UAntwerp), and Inge Bats (Institute
Born-Bunge) for administrative assistance as well as Fujirebio Europe
(Ghent, Belgium) for kindly providing the Innotest assays.
References
1. Dubois B, Feldman HH, Jacova C, Hampel H, Molinuevo
JL, Blennow K et al. Advancing research diagnostic cri-
teria for Alzheimer’s disease: the IWG-2 criteria. Lancet
Neurol 2014;13:614 –29.
2. Vanderstichele H, Bibl M, Engelborghs S, Le Bastard N,
Lewczuk P, Molinuevo JL et al. Standardization of pre-
analytical aspects of cerebrospinal fluid biomarker test-
ing for Alzheimer’s disease diagnosis: a consensus pa-
per from the Alzheimer’s Biomarkers Standardization
Initiative. Alzheimers Dement 2012;8:65–73.
3. Mattsson N, Andreasson U, Persson S, Carrillo MC, Col-
lins S, Chalbot S et al. CSF biomarker variability in the
Alzheimer’sAssociation quality control program. Alzhei-
mers Dement 2013;9:251– 61.
4. del Campo M, Mollenhauer B, Bertolotto A, Engel-
borghs S, Hampel H, Simonsen AH et al. Recommenda-
tions to standardize preanalytical confounding factors
in Alzheimer’s and Parkinson’s disease cerebrospinal
fluid biomarkers: an update. Biomark Med 2012;6:
419–30.
5. Reiber H. Dynamics of brain-derived proteins in cere-
brospinal fluid. Clin Chim Acta 2001;310:173– 86.
6. Bjerke M, Portelius E, Minthon L, Wallin A, Anckarsater
H, Anckarsater R et al. Confounding factors influencing
amyloid beta concentration in cerebrospinal fluid. Int J
Alzheimers Dis 2010;2010:986310.
7. Blennow K, Nellgard B. Amyloid beta 1–42 and tau in
cerebrospinal fluid after severe traumatic brain injury.
Neurology 2004;62:159; author reply 159 –60.
8. Tarnaris A, Toma AK, Chapman MD, Petzold A, Keir G,
Kitchen ND et al. Rostrocaudal dynamics of CSF bio-
markers. Neurochem Res 2011;36:528 –32.
9. de Bont JM, Vanderstichele H, Reddingius RE, Pieters R,
van Gool SW. Increased total-tau levels in cerebrospinal
fluid of pediatric hydrocephalus and brain tumor pa-
tients. Eur J Paediatr Neurol 2008;12:334 –41.
10. Vanderstichele H, De Vreese K, Blennow K, Andreasen
N, Sindic C, Ivanoiu A et al. Analytical performance and
clinical utility of the INNOTEST PHOSPHO-TAU
181P
as-
sayfor discrimination between Alzheimer’s disease and
dementia with Lewy bodies. Clin Chem Lab Med 2006;
44:1472–80.
11. Schoonenboom NS, Mulder C, Vanderstichele H, Van
Elk EJ, Kok A, Van Kamp GJ et al. Effects of processing
and storage conditions on amyloid beta (1– 42) and tau
concentrations in cerebrospinal fluid: implications for
use in clinical practice. Clin Chem 2005;51:189 –95.
12. Vanderstichele H, De Meyer G, Shapiro F, Engelborghs
S, De Deyn PP, Shaw LM et al. Alzheimer’s disease
biomarkers: from concept to clinical utility. Biomarkers
for Early Diagnosis of Alzheimer’s Disease. Hauppauge,
NY: Nova Science Publishers, 2008:81–122.
13. Sjögren M, Vanderstichele H, Agren H, Zachrisson O,
Edsbagge M, Wikkelso C et al. Tau and Abeta42 in ce-
rebrospinal fluid from healthy adults 21–93 years of
age: establishment of reference values. Clin Chem
2001;47:1776 81.
14. Kaiser E, Schonknecht P, Thomann PA, Hunt A, Schro-
der J. Influence of delayed CSF storage on concentra-
tions of phospho-tau protein (181), total tau protein
and beta-amyloid (1–42). Neurosci Lett 2007;417:
193–5.
15. Sancesario GM, Esposito Z, Nuccetelli M, Bernardini S,
Sorge R, Martorana A et al. Abeta1– 42 Detection in CSF
of Alzheimer’s disease is influenced by temperature:
indication of reversible Abeta1–42 aggregation? Exp
Neurol 2010;223:371– 6.
16. Lachno DR, Romeo MJ, Siemers ER, Vanderstichele H,
Coart E, Konrad RJ et al. Validation of ELISA methods for
quantification of total tau and phosporylated-tau181 in
human cerebrospinal fluid with measurement in spec-
imens from two Alzheimer’s disease studies. J Alzhei-
mers Dis 2011;26:531– 41.
17. Vanderstichele H, Van Kerschaver E, Hesse C, Da-
vidsson P, Buyse MA, Andreasen N et al. Standardiza-
tion of measurement of beta-amyloid(1–42) in cere-
brospinal fluid and plasma. Amyloid 2000;7:245–58.
18. Zimmermann R, Lelental N, Ganslandt O, Maler JM,
KornhuberJ, Lewczuk P. Preanalytical sample handling
and sample stability testing for the neurochemical de-
mentia diagnostics. J Alzheimers Dis 2011;25:739
45.
19. Sullivan JT, Grouper S, Walker MT, Parrish TB, McCarthy
RJ, Wong CA. Lumbosacral cerebrospinal fluid volume
in humans using three-dimensional magnetic reso-
nance imaging. Anesth Analg 2006;103:1306 –10.
20. Ramont L, Thoannes H, Volondat A, Chastang F, Millet
MC, Maquart FX. Effects of hemolysis and storage con-
dition on neuron-specific enolase (NSE) in cerebrospi-
nal fluid and serum: implications in clinical practice.
Clin. Chem Lab Med 2005;43:1215–7.
21. Mollenhauer B, El-Agnaf OM, Marcus K, Trenkwalder C,
Schlossmacher MG. Quantification of alpha-synuclein
in cerebrospinal fluid as a biomarker candidate: review
of the literature and considerations for future studies.
Biomark Med 2010;4:683–99.
22. You JS, Gelfanova V, Knierman MD, Witzmann FA,
Wang M, Hale JE. The impact of blood contamination
on the proteome of cerebrospinal fluid. Proteomics
2005;5:290 6.
23. Le Bastard N, Aerts L, Leurs J, Blomme W, De Deyn PP,
Engelborghs S. No correlation between time-linked
742 Clinical Chemistry 61:5 (2015)
plasma and CSF Abeta levels. Neurochem Int 2009;55:
820–5.
24. Hansson SF, Simonsen AH, Zetterberg H, Andersen O,
Haghighi S, Fagerberg I et al. Cystatin C in cerebrospi-
nal fluid and multiple sclerosis. Ann Neurol 2007;62:
193– 6; discussion 205.
25. Perret-Liaudet A, Pelpel M, Tholance Y, Dumont B,
Vanderstichele H, Zorzi W et al. Risk of Alzheimer’s dis-
ease biological misdiagnosis linked to cerebrospinal
collection tubes. J Alzheimers Dis 2012;31:13–20.
26. Biere AL, Ostaszewski B, Stimson ER, Hyman BT, Mag-
gio JE, Selkoe DJ. Amyloid beta-peptide is transported
on lipoproteins and albumin in human plasma. J Biol
Chem 1996;271:32916 –22.
27. Koudinov AR, Koudinova NV, Kumar A, Beavis RC,
Ghiso J. Biochemical characterization of Alzheimer’s
soluble amyloid beta protein in human cerebrospinal
fluid: association with high density lipoproteins.
Biochem Biophys Res Commun 1996;223:592–7.
28. Bouwman FH, van der Flier WM, Schoonenboom NS,
Van Elk EJ, Kok A, Scheltens P et al. Usefulness of
longitudinal measurements of beta-amyloid1– 42 in
cerebrospinal fluid of patients with various cognitive
and neurologic disorders. Clin Chem 2006;52:
1604 6.
29. Engelborghs S, De Vreese K, Van de Casteele T, Vander-
stichele H, Van Everbroeck B, Cras P et al. Diagnostic
performance of a CSF-biomarker panel in autopsy-
confirmed dementia. Neurobiol Aging 2008;29:
1143–59.
Preanalytical Aspects of CSF Biomarker Analyses for AD
Clinical Chemistry 61:5 (2015) 743

Supplementary resource (1)

... It is not clear precisely which confounding factor(s) might contribute to the variance observed in the current study. No gradient effect was observed in CSF by Le Bastard et al. (2015) for Aβ42, total tau, and pTau181P as measured by ELISA (n = 20) [37]. Some gradient effect was seen with higher collection volumes for CSF Aβ42 when analysed on an automated chemiluminesent platform [38]. ...
... It is not clear precisely which confounding factor(s) might contribute to the variance observed in the current study. No gradient effect was observed in CSF by Le Bastard et al. (2015) for Aβ42, total tau, and pTau181P as measured by ELISA (n = 20) [37]. Some gradient effect was seen with higher collection volumes for CSF Aβ42 when analysed on an automated chemiluminesent platform [38]. ...
Article
Full-text available
Background CSF biomarkers are well-established for routine clinical use, yet a paucity of comparative assessment exists regarding CSF extraction methods during lumbar puncture. Here, we compare in detail biomarker profiles in CSF extracted using either gravity drip or aspiration. Methods Biomarkers for β-amyloidopathy (Aβ1–42, Aβ1–40), tauopathy (total tau), or synapse pathology (BACE1, Neurogranin Trunc-p75, α-synuclein) were assessed between gravity or aspiration extraction methods in a sub-population of the Australian Imaging, Biomarkers and Lifestyle (AIBL) study (cognitively normal, N = 36; mild cognitive impairment, N = 8; Alzheimer’s disease, N = 6). Results High biomarker concordance between extraction methods was seen (concordance correlation > 0.85). Passing Bablock regression defined low beta coefficients indicating high scalability. Conclusions Levels of these commonly assessed CSF biomarkers are not influenced by extraction method. Results of this study should be incorporated into new consensus guidelines for CSF collection, storage, and analysis of biomarkers.
... For a long time, manual enzyme-linked immunosorbent assay (ELISA) has been widely employed as the reference method for the analysis of the CSF AD biomarkers. However, its broad-scale use is critically hampered by the assay variability, which influences the measurement of the analytes and the interpretation of the outcome data, especially in the routine clinical context Le Bastard et al., 2015). Due to these concerns, ELISA was recently replaced in worldwide laboratories by fully automated assays -such as chemiluminescence enzyme immunoassay (CLEIA) -which offer grounds to cut sample manipulation steps and to reduce the intra-and inter-laboratory variability for CSF biomarker measurement (Kollhoff et al., 2018). ...
... Hierarchical clustering was used for ordering diagnostic groups and clusters. respect to manual ELISA (Le Bastard et al., 2015). In this work, we used the reproducibility of automated CSF core AD biomarker determination to compare two large cohorts of patients diagnosed with various neurological disorders and enrolled in two centers with proven expertise in AD biomarkers. ...
Article
Full-text available
Amyloid-beta (Ab) 42/40 ratio, tau phosphorylated at threonine-181 (p-tau), and totaltau (t-tau) are considered core biomarkers for the diagnosis of Alzheimer’s disease (AD). The use of fully automated biomarker assays has been shown to reduce the intraand inter-laboratory variability, which is a critical factor when defining cut-off values. The calculation of cut-off values is often influenced by the composition of AD and control groups. Indeed, the clinically defined AD group may include patients affected by other forms of dementia, while the control group is often very heterogeneous due to the inclusion of subjects diagnosed with other neurological diseases (OND). In this context, unsupervised machine learning approaches may overcome these issues providing unbiased cut-off values and data-driven patient stratification according to the sole distribution of biomarkers. In this work, we took advantage of the reproducibility of automated determination of the CSF core AD biomarkers to compare two large cohorts of patients diagnosed with different neurological disorders and enrolled in two centers with established expertise in AD biomarkers. We applied an unsupervised Gaussian mixture model clustering algorithm and found that our large series of patients could be classified in six clusters according to their CSF biomarker profile, some presenting a typical AD-like profile and some a non-AD profile. By considering the frequencies of clinically defined OND and AD subjects in clusters, we subsequently computed cluster-based cut-off values for Ab42/Ab40, p-tau, and t-tau. This approach promises to be useful for large-scale biomarker studies aimed at providing efficient biochemical phenotyping of neurological diseases.
... Pre-analytical factors like specimen collection, hemolytic contamination of samples, sample handling, storage conditions and temperature, sample stability over freeze-thaw prior to processing, and batch-to batch variability of the kits utilized are critical methodological issues that need consideration (see reviews [121][122][123][124][125]). The volume of CSF collected can influence the protein concentrations, although the spinal cord gradient effect on Aβ42 has not been demonstrated-though it is suggested to standardize the volume of CSF withdrawal to 12 mL [126]. ...
Article
Full-text available
Alzheimer’s disease (AD) is an irreversible neurodegenerative disease characterized by progressive cognitive decline. The two cardinal neuropathological hallmarks of AD include the buildup of cerebral β amyloid (Aβ) plaques and neurofibrillary tangles of hyperphosphorylated tau. The current disease-modifying treatments are still not effective enough to lower the rate of cognitive decline. There is an urgent need to identify early detection and disease progression biomarkers that can facilitate AD drug development. The current established readouts based on the expression levels of amyloid beta, tau, and phospho-tau have shown many discrepancies in patient samples when linked to disease progression. There is an urgent need to identify diagnostic and disease progression biomarkers from blood, cerebrospinal fluid (CSF), or other biofluids that can facilitate the early detection of the disease and provide pharmacodynamic readouts for new drugs being tested in clinical trials. Advances in proteomic approaches using state-of-the-art mass spectrometry are now being increasingly applied to study AD disease mechanisms and identify drug targets and novel disease biomarkers. In this report, we describe the application of quantitative proteomic approaches for understanding AD pathophysiology, summarize the current knowledge gained from proteomic investigations of AD, and discuss the development and validation of new predictive and diagnostic disease biomarkers.
... Another source of noise that can greatly affect biomarker values is systematic error caused by assay-related changes in the analytical performance of the methods such as when changing lots of key materials (such as antibodies or calibrators). 27,28 However, systematic error is much more difficult to quantify empirically given its unpredictable nature. Systematic error can also greatly degrade performance of diagnostic models when the assay is characterized by low dynamic range or when there is high overlap between positive and negative groups. ...
Article
Full-text available
Introduction: The effect of random error on the performance of blood-based biomarkers for Alzheimer's disease (AD) must be determined before clinical implementation. Methods: We measured test-retest variability of plasma amyloid beta (Aβ)42/Aβ40, neurofilament light (NfL), glial fibrillary acidic protein (GFAP), and phosphorylated tau (p-tau)217 and simulated effects of this variability on biomarker performance when predicting either cerebrospinal fluid (CSF) Aβ status or conversion to AD dementia in 399 non-demented participants with cognitive symptoms. Results: Clinical performance was highest when combining all biomarkers. Among single-biomarkers, p-tau217 performed best. Test-retest variability ranged from 4.1% (Aβ42/Aβ40) to 25% (GFAP). This variability reduced the performance of the biomarkers (≈ΔAUC [area under the curve] -1% to -4%) with the least effects on models with p-tau217. The percent of individuals with unstable predicted outcomes was lowest for the multi-biomarker combination (14%). Discussion: Clinical prediction models combining plasma biomarkers-particularly p-tau217-exhibit high performance and are less effected by random error. Individuals with unstable predicted outcomes ("gray zone") should be recommended for further tests.
... Considering the strong influence of pre-analytical conditions on biomarkers evaluation [6,9,[13][14][15]17,25], we ensured the harmonization of all these parameters in our study. Thus, for a same patient's CSF, the storage volume (and therefore the surface/volume ratio) was the only variable that changed through experiments. ...
Article
Objectives Alzheimer’s disease (AD) is considered the most common cause of dementia in older people. Cerebrospinal fluid (CSF) Aβ1-42, Aβ1-40, total Tau (t-Tau), and phospho Tau (p-Tau) are important biomarkers for the diagnosis, however, they are highly dependent on the pre-analytical conditions. Our aim was to investigate the potential influence of different storage conditions on the simultaneous quantification of these biomarkers in a fully-automated platform to accommodate easier pre-analytical conditions for laboratories. Methods CSF samples were obtained from 11 consecutive patients. Aβ1-42, Aβ1-40, p-Tau, and t-Tau were quantified using the LUMIPULSE G600II automated platform. Results Temperature and storage days significantly influenced Aβ1-42 and Aβ1-40 with concentrations decreasing with days spent at 4 °C. The use of the Aβ1-42/Aβ1-40 ratio could partly compensate it. P-Tau and t-Tau were not affected by any of the tested storage conditions. For conditions involving storage at 4 °C, a correction factor of 1.081 can be applied. Diagnostic agreement was almost perfect in all conditions. Conclusions Cutoffs calculated in samples stored at −80 °C can be safely used in samples stored at −20 °C for 15–16 days or up to two days at RT and subsequent freezing at −80 °C. For samples stored at 4 °C, cutoffs would require applying a correction factor, allowing to work with the certainty of reaching the same clinical diagnosis.
... However, similar to blood, discrepancies in the handling of CSF samples may result in preanalytical variation of protein measures 8,9 . For example, CXCL13, a biomarker for multiple sclerosis (MS) 10 , can be affected by the frequency of freeze/thaw cycles 11 , while β-amyloid and total tau levels used for the diagnosis of Alzheimer's disease can also be influenced by delays in sample processing 9,12 . ...
Article
Full-text available
Proteomics studies are important for the discovery of new biomarkers as clinical tools for diagnosis and disease monitoring. However, preanalytical variations caused by differences in sample handling protocol pose challenges for assessing biomarker reliability and comparability between studies. The purpose of this study was to examine the effects of delayed centrifuging on measured protein levels in plasma and CSF. Blood from healthy individuals and patients with multiple sclerosis along with CSF from patients with suspected neurological disorders were left at room temperature for different time periods (blood: 1, 24, 48, 72 hours; CSF: 1 and 6 hours) prior to centrifuging. Ninety-one inflammation-related proteins were analyzed using a proximity extension assay, a high sensitivity multiplex immunoassay. Additional metabolic and neurology related markers were also investigated in CSF. In summary, many proteins, particularly in plasma, had increased levels with longer delays in processing likely due in part to intracellular leakage. Levels of CASP-8, IL-8, IL-18, SIRT2, and ST1A1 increased 2-10 fold in plasma after 24 hours at room temperature. Similarly, levels of CTSH, ENTPD5, and WWP2 differentiated in CSF with <6 hour delay in processing. However, the rate of change for many proteins was relatively consistent, therefore, we were able to characterize biomarkers for detecting sample handling variability. Our findings highlight the importance of timely and consistent sample collection and the need for increased awareness of protein susceptibility to sample handling bias. In addition, suggested biomarkers may be used in certain situations to detect and correct for preanalytical variation in future studies.
... It is known that there is a variability in analytical results of AD biomarkers [43,44] and preanalytical factors, such as CSF collection, storage, and adsorption to tube-walls, have been shown to contribute the highest to this variability [45]. Concerning storage preanalytical variability, the CSF AD biomarkers are stable in -80 • C storage [27], whereas freezing at -20 • C seems to affect concentrations of different CSF biomarkers [27,46]. We present a simple and validated ELISA assay that enables retrospective characterization of CSF biospecimens without the use of mass spectrometry. ...
Article
Full-text available
Background: An N-terminal octapeptide cleavage of the cystatin C protein was discovered by mass spectrometry when cerebrospinal fluid (CSF) was stored at -20°C for 3 months, which did not occur when CSF was stored at -80°C. Objective: The aim was to develop an immunoassay as quality assessment tool to detect this -20°C cleavage of cystatin C in CSF and support Alzheimer's disease research. Methods: A specific monoclonal antibody and a double indirect sandwich ELISA were developed: one assay quantifies the octapeptide uncleaved protein specifically and the other quantifies the total cystatin C present in the biological fluid (both cleaved and uncleaved forms). The ratio of these concentrations was calculated to assess the extent of cleavage of cystatin C. The novel ELISA was validated and applied in a short-term (up to 4 weeks) and mid-term (up to one year) stability study of CSF stored at 4°C, -20°C, -80°C, and liquid nitrogen. Impact of freeze-thaw cycles, adsorption, and protease inhibitors were tested. Results: The ratio of truncated protein was modified following -20°C storage and seemed to reach a plateau after 6 months. The ratio was impacted neither by freeze-thaw cycles nor adsorption. The -20°C specific cleavage was found to be protease related. Conclusion: Using this novel double indirect sandwich ELISA, absolute levels of the total and uncleaved cystatin C and the ratio of truncated cystatin C can be measured. This assay is an easily applicable tool which can be used to confirm that CSF biospecimen are fit-for-purpose for Alzheimer's disease research.
... Enfin acheminement, traitement et conservation avant analyse doivent également respecter certaines règles : les délais maximum d'acheminement au laboratoire local dépendent de la température d'acheminement mais aussi du tube de prélèvement utilisé : ils sont de 24 heures à température ambiante et de 72 h à une semaine à 4 8C ; la centrifugation des échantillons permet d'éliminer les cellules sanguines en cas de contamination sanguine légère de l'échantillon de LCS (< 5000 GR/mL). Les échantillons sont ensuite conservés à À80 8C jusqu'à analyse [24]. Le transport à À20 8C des aliquots congelés vers le laboratoire central réalisant les dosages est possible. ...
Article
Full-text available
Résumé Le dosage des biomarqueurs du liquide cérébrospinal (LCS) a intégré depuis maintenant plusieurs années la démarche diagnostique en cas de suspicion de maladie d’Alzheimer (MA). En France, où l’imagerie moléculaire n’est pas accessible en soin courant, ils sont le seul outil disponible pour poser un diagnostic biologique de MA selon les critères de 2018. La combinaison de Aβ1-42 et Aβ1-40 prédit avec une excellente sensibilité et spécificité la présente d’une pathologie amyloïde (A), tandis que P-Tau et T-tau sont des biomarqueurs de la taupathie associée à la MA (T), T-tau ayant une moindre spécificité que P-tau. L’association de ces quatre biomarqueurs peut différencier une MA d’une autre maladie neurodégénérative et poser un diagnostic de MA dès le stade prodromal ou préclinique. Les biomarqueurs ont toutefois des limites qui tiennent d’abord à la variabilité du dosage, liée à des facteurs pré-analytiques et analytiques. L’harmonisation des procédures, l’automatisation et les efforts encore en cours de standardisation contribuent à réduire les sources de variabilité. Leur interprétation est souvent difficile, liée notamment aux profils ambigus (A+T- ou A-T+) et à la fréquence des associations pathologiques, notamment chez le sujet très âgé. Ils conservent néanmoins à tout âge une excellente valeur prédictive négative à tout âge. L’arrivée annoncée des biomarqueurs sanguins relèguera peut-être les biomarqueurs du LCS au second plan. Ils resteront néanmoins pour les années à venir un gold standard biologique.
Article
Background: RNA integrity of tumor tissues from 12 common organs was measured, and tumor tissues from liver were found to have the best RNA integrity in our previous study. The effects of preanalytical variables in the phase of pre- and postacquisition on RNA integrity were further assessed in hepatocellular carcinoma (HCC) tissues in this study. Methods: RNA integrity number (RIN) was measured in tissues from 146 HCC patients. First, 42 fresh HCC tumor tissues were newly collected to assess the effect of various preanalytical variables in the phase of preacquisition on RNA integrity. Second, eight paired HCC tumor and normal tissues were newly collected and used in the gradient course study of ex vivo ischemia time and freeze-thaw cycles on RNA integrity. Finally, 96 stock-frozen tumor tissues with various years of frozen storage were used to assess the effect of cryopreservation time. Results: RNA integrity was found to be independent of patient age, sex, clinical stage, tumor location, HBV infection status, tumor diameter, and surgical approach, but affected by tumor grade. Tumor tissues with a greater tumor grade had lower RIN. With the prolongation of ex vivo ischemia time, freeze-thaw cycles, and cryopreservation time, the RIN of HCC tissues showed decreasing trends. Significant decreases in RIN of the tumor and normal tissues were observed at 6 and 2 hours of ex vivo ischemia time, respectively, and significantly decreased RIN of tumor tissues was observed after six freeze-thaw cycles and 6 years of cryopreservation. Conclusions: Preanalytical variables in the phase of preacquisition such as tumor grade, and in the postacquisition phase such as ex vivo ischemia time, freeze-thaw times, and freeze-storage time both have effects on RNA integrity of HCC tissues. Tissue-based translational research should pay attention to preanalytical variables when collecting and utilizing tumor tissues.
Article
This study investigated the effect of sample storage duration on the quantification of oxidative stress markers in the gastrocnemius, heart, and brain of mice submitted to a maximum swimming exercise. Thiobarbituric acid reactive substances (TBARSs), protein carbonyl derivatives, total antioxidant capacity (TAC), and the activity of superoxide dismutase (SOD) and catalase (CAT) were quantified in fresh tissues and in samples stored at -80°C for 1, 3, or 6 months, from exercised (n = 13) and nonexercised mice (n = 13). Except for protein carbonyl derivatives in the heart, the exercise resulted in the modification of all markers in all fresh-evaluated samples (p < 0.001). The storage duration did not modify the effect of exercise on protein carbonyl derivatives and TAC. TBARS was stable for 3 months in the gastrocnemius and for 1 month in frozen heart and brain. Accordingly, the exercise effect on TBARS levels observed in fresh samples was absent in the gastrocnemius frozen for 6 months (p = 0.98) and in the heart and brain frozen for 3 months (p = 0.07 and 0.28, respectively) or more (p = 0.21 for heart and p > 0.99 for brain). In addition, CAT and SOD activities were reduced by storage duration in all tissues evaluated (p < 0.05). Our findings show that sample storage duration alters the quantification of oxidative stress markers in mice submitted to maximum exercise, and its effect is tissue and marker dependent. Some recommendations to achieve more accurate and reproducible data in the exercise physiology and oxidative stress markers field are presented.
Article
Full-text available
Background: Tau protein and the 42-amino acid form of beta -amyloid (A beta 42) measured in cerebrospinal fluid (CSF) have been proposed as potential biochemical diagnostic markers for Alzheimer disease. For the introduction of these assays in clinical practice, adequate reference values are of importance. Methods: CSF samples were obtained from 231 neurologically and psychiatrically healthy individuals, 21-93 years of age, all with a MiniMental State examination score of 28 or above. Standardized ELISAs were used to measure tau and A beta 42 in CSF. Following IFCC recommendations, we used a rank-based method; the 0.90 and 0.10 fractiles were estimated to establish reference values for CSF-tau and CSF-A beta 42, respectively. Putative confounding factors, such as the influence of the passage of proteins from peripheral blood to CSF, influence of dysfunction of the blood-brain barrier, and freezing and thawing of CSF, were investigated. Results: A correlation with age was found for CSF-tau (r = 0.60; P <0.001). Therefore, separate reference values for different age groups were established for CSF-tau: <300 ng/L in the group 21-50 years of age, <450 ng/L in the group 51-70 years of age, and < 500 ng/L in the group 71-93 years of age. CSF-A beta 42 did not correlate with age (r = -0.045), and the reference value was set to >500 ng/L. No correlation was found between blood-brain barrier function and CSF-tau or CSF-A beta 42. Conclusions: These reference values can be applied when using CSF-tau and CSF-A beta 42 in clinical practice. (C) 2001 American Association for Clinical Chemistry.
Article
Full-text available
In the past 8 years, both the International Working Group (IWG) and the US National Institute on Aging-Alzheimer's Association have contributed criteria for the diagnosis of Alzheimer's disease (AD) that better define clinical phenotypes and integrate biomarkers into the diagnostic process, covering the full staging of the disease. This Position Paper considers the strengths and limitations of the IWG research diagnostic criteria and proposes advances to improve the diagnostic framework. On the basis of these refinements, the diagnosis of AD can be simplified, requiring the presence of an appropriate clinical AD phenotype (typical or atypical) and a pathophysiological biomarker consistent with the presence of Alzheimer's pathology. We propose that downstream topographical biomarkers of the disease, such as volumetric MRI and fluorodeoxyglucose PET, might better serve in the measurement and monitoring of the course of disease. This paper also elaborates on the specific diagnostic criteria for atypical forms of AD, for mixed AD, and for the preclinical states of AD.
Article
Full-text available
The standardization and clinical validation of the measurement of β–amyloid. (Aβ42) in cerebrospinal fluid (CSF), plasma and urine is described using a commercially available sandwich–type ELISA with 2 IF12 and 3D6 as monoclonal antibodies. The INNOTEST™ β–amyloid(1–42) allows the specific and reliable measurement of (1–42) amyloid peptides in CSF and plasma. The Aβ42 concentrations in serum and urine were below the detection limit. In plasma, no differences were found in Aβ42 levels between controls and patients with different neurodegenerative disorders (Alzheimer's disease (AD), Lewy body disease (LED), others). In contrast, CSF–Aβ42 concentrations were lower in AD and LBD patients as compared to controls. No correlation was found in AD patients between CSF and plasma concentrations of Aβ42 or between CSF Aβ42 levels and blood–brain–barrier function. The quantitative outcome of the test is in part dependent on confounding factors such as tube type, freeze/thaw cycles, temperature of incubation, standard preparation protocol, and antibody selection. Notwithstanding these aspects, it emerged that Aβ42 is a useful biochemical marker for the diagnosis of AD patients, but there is a need for an international Aβ standard, a universally accepted protocol for CSF preparation, and a thorough evaluation of assay performance in function of the boundary conditions.
Article
Full-text available
Early diagnosis of neurodegenerative disorders such as Alzheimer's (AD) or Parkinson's disease (PD) is needed to slow down or halt the disease at the earliest stage. Cerebrospinal fluid (CSF) biomarkers can be a good tool for early diagnosis. However, their use in clinical practice is challenging due to the high variability found between centers in the concentrations of both AD CSF biomarkers (Aβ42, total tau and phosphorylated tau) and PD CSF biomarker (α-synuclein). Such a variability has been partially attributed to different preanalytical procedures between laboratories, thus highlighting the need to establish standardized operating procedures. Here, we merge two previous consensus guidelines for preanalytical confounding factors in order to achieve one exhaustive guideline updated with new evidence for Aβ42, total tau and phosphorylated tau, and α-synuclein. The proposed standardized operating procedures are applicable not only to novel CSF biomarkers in AD and PD, but also to biomarkers for other neurodegenerative disorders.
Article
Full-text available
Tau proteins and amyloid-β (Aβ) peptides are the current recognized cerebrospinal fluid (CSF) biomarkers used as an aid in the diagnosis of Alzheimer's disease (AD). However, there is no consensus on their clinical use due to non-qualified cut-off values, probably related to the observed high pre-analytical and analytical variability. Standardized pre-analytical protocols have therefore been proposed. Importantly, these recommend the use of polypropylene collection/sampling tubes while, to date, no broad comparison of these types of tubes has been conducted. In this study, we first compared, as part of a real clinical workflow, the impact of four different collection tubes on the CSF concentration of Aβ peptides (Aβ42, Aβ40) and total (hTau) and phosphorylated (P-Tau181P) tau proteins measured using routine ELISA kits. We then extended this study to 11 polypropylene tubes used by different clinical laboratories, and investigated their plastic polymer composition using differential scanning calorimetry and Fourier Transformed Infrared spectroscopy. Significant concentration variations linked solely to the use of different types of tubes were observed. This was particularly marked for Aβ peptides, with >50% disparity occurring in less than five minutes. Polymer composition analysis revealed that most polypropylene tubes were in fact copolymers with at least polyethylene. There was no clear correlation between tube composition and pre-analytical behavior. Our results show that the use of polypropylene tubes does not guarantee satisfactory pre-analytical behavior. They also point to collection/sampling tubes being a major pre-analytical source of variability that could impact the significance of AD biological diagnosis.
Article
The cerebrospinal fluid (CSF) biomarkers amyloid beta 1-42, total tau, and phosphorylated tau are used increasingly for Alzheimer's disease (AD) research and patient management. However, there are large variations in biomarker measurements among and within laboratories. Data from the first nine rounds of the Alzheimer's Association quality control program was used to define the extent and sources of analytical variability. In each round, three CSF samples prepared at the Clinical Neurochemistry Laboratory (Mölndal, Sweden) were analyzed by single-analyte enzyme-linked immunosorbent assay (ELISA), a multiplexing xMAP assay, or an immunoassay with electrochemoluminescence detection. A total of 84 laboratories participated. Coefficients of variation (CVs) between laboratories were around 20% to 30%; within-run CVs, less than 5% to 10%; and longitudinal within-laboratory CVs, 5% to 19%. Interestingly, longitudinal within-laboratory CV differed between biomarkers at individual laboratories, suggesting that a component of it was assay dependent. Variability between kit lots and between laboratories both had a major influence on amyloid beta 1-42 measurements, but for total tau and phosphorylated tau, between-kit lot effects were much less than between-laboratory effects. Despite the measurement variability, the between-laboratory consistency in classification of samples (using prehoc-derived cutoffs for AD) was high (>90% in 15 of 18 samples for ELISA and in 12 of 18 samples for xMAP). The overall variability remains too high to allow assignment of universal biomarker cutoff values for a specific intended use. Each laboratory must ensure longitudinal stability in its measurements and use internally qualified cutoff levels. Further standardization of laboratory procedures and improvement of kit performance will likely increase the usefulness of CSF AD biomarkers for researchers and clinicians.
Article
Numerous studies show that the cerebrospinal fluid biomarkers total tau (T-tau), tau phosphorylated at threonine 181 (P-tau(181P)), and amyloid-β (1-42) (Aβ(1-42)) have high diagnostic accuracy for Alzheimer's disease. Variability in concentrations for Aβ(1-42), T-tau, and P-tau(181P) drives the need for standardization. Key issues were identified and discussed before the first meeting of the members of the Alzheimer's Biomarkers Standardization Initiative (ABSI). Subsequent ABSI consensus meetings focused on preanalytical issues. Consensus was reached on preanalytical issues such as the effects of fasting, different tube types, centrifugation, time and temperature before storage, storage temperature, repeated freeze/thaw cycles, and length of storage on concentrations of Aβ(1-42), T-tau, and P-tau(181P) in cerebrospinal fluid. The consensus reached on preanalytical issues and the recommendations put forward during the ABSI consensus meetings are presented in this paper.