ORIGINAL ARTICLE INTERNATIONAL JOURNAL OF LA BO RATO RY HEMATOLOGY

Evaluation and performance characteristics of the

coagulation system: ACL TOP analyzer – HemosIL reagents
M. MILOS*, D. HERAK*, L. KURIC †, I. HORVAT †, R. ZADRO*

*
Clinical Institute of Laboratory
Diagnosis, Clinical Hospital Center
Zagreb University School of
Medicine, Zagreb, Croatia,
†
Faculty of Pharmacy and
Biochemistry, University of Zagreb,
Zagreb, Croatia
Correspondence:
Dr Renata Zadro, Clinical Institute
of Laboratory Diagnosis, Zagreb
Clinical Hospital Center, Kispaticeva
12, 10000 Zagreb, Croatia.
Tel.: +385 1 2388 247;
Fax: +385 1 2312 079;
E-mail: rzadro@mef.hr
doi:10.1111/j.1751-553X.2007.00999.x
Received 2 May 2007; accepted for
publication 2 October 2007
Keywords
Coagulation analyzer, ACL TOP,
HemosIL reagents, evaluation
SUMMARY
ACL TOP is a fully automated coagulation analyzer, designed for
simultaneous measurement of routine and special coagulation param-
eters. We evaluated analytical and technical performance characteris-
tics of the coagulation system composed of the ACL TOP analyzer and
HemosIL reagent group for the determination of routine clotting (PT,
APTT, fibrinogen, FVII, and FVIII), chromogenic (protein C) and
immunological assays (FXIII antigen). Within run and between run
CVs ranged from 0.9% to 7.7% and from 2.0% to 14.8% respectively.
The obtained CVs for imprecision of calibration curves were <5% for PT
and <7% for fibrinogen. The method comparison study showed good
correlation between results obtained on the ACL TOP and BCS/BCT
analyzers, with correlation co-efficients ranging from 0.709 to 0.955,
but with significantly different results for PT INR, APTT, fibrinogen and
protein C, and wide dispersion of differences observed in difference
plots for most assays. Despite good correlation and agreement for FVIII,
problems in measuring FVIII<10% were encountered. The effective
throughput for the ACL TOP and BCS was 151 and 212 PT/APTT/
fibrinogen tests per hour, respectively. Although the ACL TOP
is designed to run multiple assays on a large number of samples,
software limitations make the instrument suitable rather for mid-sized
laboratories.

designed for the screening of a large number of sam-

INTRODUCTION
During the last 15 years, coagulation analyzers and
reagents have changed significantly. The current gener-
ation of coagulation analyzers is fully automated, and
combines several technologies in the same instrument
(Walenga & Fareed, 1994). By using clotting, chromo-
genic, and immunological methods for quantification
of basic and specific hemostatic parameters, they are
26 Journal compilation Ó 2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2009, 31, 26–35
ples, with capabilities for performing a wide range of
different coagulation assays on a single or on multiple
samples at the same time. Along with the technological
development of analyzers, major improvement has
been made in software support, offering users easier
and more efficient processing of daily routine.
The aim of the present study was to evaluate ana-
lytical and technical performance of a coagulation
Ó 2007 The Authors
 M. MILOS ET AL.
testing system composed of the ACL TOP coagulation
analyzer and HemosIL reagent group (Instrumentation
Laboratory, Lexington, MA, USA) under routine labo-
ratory conditions. We tested prothrombin time (PT),
activated partial thromboplastin time (APTT), fibrino-
gen (Fib), factor VII (FVII), factor VIII (FVIII) and pro-
tein C (PC) activities, and FXIII antigen (FXIII : Ag).
The evaluation addressed several issues, including
available methodologies, validation of performance,
throughput study, reagent and patient sample on-
board capabilities and ease of operation.
MATERIALS AND METHODS
Description of the ACL TOP analyzer
The ACL TOP analyzer is a fully automated, random-
access, multiparameter coagulation analyzer designed
for simultaneous measurement of routine and special
coagulation parameters. The system is composed of
control and analytical modules. The analytical module
consists of the functionality required to handle and
process samples, reagents and auxiliary materials, and
can perform clotting, chromogenic and latex immuno-
assay methods. Sample and reagent areas have covers
that provide increased operating safety and reduce the
influence of external environment. The ACL TOP
allows continuous loading of cuvettes, reagents and
samples. The cuvette loading area is on the left side of
the instrument, allowing the maximum single load of
800 cuvettes. The sample area is at room temperature
and can hold 12 racks, each capable of holding 10
samples either in primary sample tubes or in original
sample cups. The reagent area is on the right side of
the analytical module and has 60 positions for
reagents, with 44 positions cooled at 15 ± 3 °C, and
16 positions at room temperature dedicated for con-
trols, calibrators, and diluents. Samples and reagents
are delivered to the instrument on sample and reagent
racks, with barcode identification. The sample arm
consists of a heated probe and syringe used for aspi-
rating and dispensing samples. There are two reagent
arms: the left arm for aspirating and dispensing
diluents and intermediate reagents and the right arm
for aspirating and dispensing start reagents. Each
probe has a level sensor. Optical reading is performed
using two wavelengths: 405 or 671 nm for clotting
assays, 405 nm for chromogenic, and 405 or 671 nm
Ó 2007 The Authors
Journal compilation Ó 2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2009, 31, 26–35
EVALUATION OF ACL TOP – HEMOSIL REAGENT SYSTEM 27
for immunologic measurements. The control module
consists of the software running under Windows 2000
Operating System (version 2.1) (Microsoft Corpora-
tion, Redmond, WA, USA) and touch-screen. The
analyzer has the software capabilities for automatic
rerun, automatic reflex testing, STAT sample loading,
and factor parallelism testing. There is a large results
database (20 000 samples with up to 30 assays per
sample) and a possibility of access to reaction curves.
Daily maintenance is a semiautomatic procedure and
usually lasts approximately 7 min.
Evaluated samples
Patient samples referred to the laboratory for routine
coagulation testing were used for evaluation proce-
dure. Platelet-poor plasma was prepared by centrifu-
gation at 2000 g for 15 min at room temperature from
blood specimens collected into siliconized glass tubes
(Vacutainer Becton Dickinson, Meylan Cedex, France)
containing 0.105 mol/l buffered sodium citrate. Plasma
samples were used for precision, correlation and
throughput study. Commercially available lyophilized
plasma samples from Instrumentation Laboratory: He-
mosIL Normal Control, HemosIL Low Abnormal Con-
trol and Special Test Level 2 were used for quality
control and for precision study.
Reagents and assay procedures
The reagents and the calibrators (brand name and
manufacturer) used on all analyzers are listed in
Table 1, as well as test methodology. All reagents, cal-
ibrators, controls and diluents used on the ACL TOP
were from HemosIL group of reagents. Assays were
performed according to the manufacturer’s specifica-
tions and the standard laboratory methods. Calibra-
tion plasma (HemosIL Calibration plasma) was used
and diluted automatically by the analyzer to create
specific calibration curves.
PT was initiated by adding thromboplastin solu-
tion (RecombiPlasTin) to plasma sample, without
incubation.
For APTT, two different reagents were evaluated
on the ACL TOP: SynthASil (APTT1) and APTT-SP
reagent (APTT2), both with colloidal silica as an
activator. A mixture of APTT reagent and plasma sam-
ple was incubated at 37 °C for 300 s, and calcium
28 M. MILOS ET AL. EVALUATION OF ACL TOP – HEMOSIL REAGENT SYSTEM

Table 1. Reagents and calibrators used on the ACL TOP, BCS and BCT

ACL TOP reagent/calibrator BCS reagent/calibrator BCT reagent/calibrator

Test Methodology Instrumentation Laboratory Dade Behring Dade Behring

PT Coagulation RecombiPlasTin Innovin –

HemosIL Calibration plasma PT-Multi Calibrator
APTT1 Coagulation SynthASil Actin FS –
APTT2 Coagulation APTT-SP Actin FS –
Fib Coagulation FibrinogenC-XL Multifibren U –
HemosIL Calibration plasma Fibrinogen Standards 1–6
PC Chromogenic Protein C Berichrom Protein C –
HemosIL Calibration plasma Standard Human Plasma
FVII Coagulation FVII deficient plasma – FVII deficient plasma
RecombiPlasTin Innovin
HemosIL Calibration plasma Standard Human Plasma
FVIII Coagulation FVIII deficient plasma – FVIII deficient plasma
SynthASil Actin FS
HemosIL Calibration plasma Standard Human Plasma
FXIII : Ag Immunological FXIII Antigen – –
HemosIL Calibration plasma
FXIII : Act Chromogenic – Berichrom FXIII –
Standard Human Plasma

chloride (CaCl2) solution was added (0.025 mol/l
CaCl2 for APTT-SP and 0.02 mol/l CaCl2 for
SynthASil).
Fib was measured according to Clauss (1957) by
adding plasma sample, diluted 1 : 10 in Factor Diluent
solution, to thrombin solution (35 UNIH/ml) (Fibrino-
gen C-XL). Samples with Fib concentration >6.5 and
<0.8 g/l were remeasured using Fib-high and Fib-low
assay procedures, respectively.
PC activity was determined using a chromogenic
assay by incubating plasma sample, diluted 1 : 4 in
Protein C Diluent solution, and PC activator solution
(fraction from the venom of Agkistrodon contortrix con-
tortrix, 0.4 U/ml) for 270 s. After adding PC Chromo-
genic substrate S2366 (1.5 mg/ml), the change in
OD405 nm was recorded.
For FVII activity measurement, FVII deficient
plasma was added to plasma sample, diluted 1 : 10 in
Factor Diluent solution. After incubation at 37 °C,
final measurement was performed after the addition
of RecombiPlasTin.
For FVIII activity measurement, FVIII deficient
plasma was mixed with plasma sample, diluted 1 : 10
in Factor Diluent solution, and SynthASil. After incu-
bation at 37 °C, final measurement was performed
Journal compilation Ó 2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2009, 31, 26–35
after adding of the 0.02 mol/l CaCl2 solution. The
analyzer automatically remeasured all samples with
FVIII activities <10%, using different dilution of
plasma and factor diluent on the same calibration
curve.
For FXIII : Ag determination, plasma sample,
diluted 1 : 10 in FXIII Diluent solution, was preincu-
bated at 37 °C. After adding of FXIII Buffer, and incu-
bation, FXIII Latex reagent (latex polystyrene particles
coated with a rabbit polyclonal antibody against the
A-subunit of FXIII) was added and the change in
OD671 nm was recorded.
Measured clotting times of coagulation methods
(PT, Fib, FVII, and FVIII) or optical densities in chro-
mogenic (PC) or immunological (FXIII : Ag) methods
were automatically converted into final results of spe-
cific activities or antigen concentrations by using the
appropriate calibration curve. International normal-
ized ratio (INR) was calculated by the analyzer using
international sensitivity index value (ISI = 0.800) sta-
ted for the ACL TOP by the manufacturer. An auto-
matic rerun test with different sample dilution,
preprogrammed by the manufacturer, was used in
cases of levels above or below the measuring limits of
calibration curves.
Ó 2007 The Authors
 M. MILOS ET AL. EVALUATION OF ACL TOP – HEMOSIL REAGENT SYSTEM 29

of rank correlation were used. The agreement

Evaluation procedure
between results obtained on different analyzers (ACL
Evaluation procedure included determination of TOP vs. BCS/BCT) was evaluated according to Bland
within run precision, between run precision, accuracy and Altman (1986).
and correlation study for all tested parameters (PT,
APTT1 and APTT2, Fib, PC, FVII, FVIII, and
RESULTS
FXIII : Ag), precision of calibration curves for PT and
Fib, and throughput study for global assays (PT,
Within run precision
APTT1, and Fib). Within run precision was evaluated
by measuring tested parameters in normal and abnor- The results of within run precision are shown in
mal plasma samples consecutively 20 times (PT, Table 2. CVs obtained for most assays were within the
APTT1, and APTT2, Fib) or 10 times (PC, FVII, FVIII, criteria for acceptance.
and FXIII : Ag), in the same run. Between run preci-
sion was determined by repeated testing of commer-
Between run precision and accuracy
cial normal (HemosIL Normal Control) and abnormal
control plasma samples (HemosIL Low Abnormal Con- As shown in Table 3, CVs for most assays were within
trol, Special Test Level 2) over a period of 10 days. the criteria for acceptance, except for PT in HemosIL
Accuracy was evaluated by calculating the biases from Normal Control (CV = 7.7%) and PC and FVIII in
the target values stated for commercial control plasma Special Test Level 2 (CV = 14.8% and 13.7%, respec-
samples for all assays. For calibration curve precision tively). In the accuracy study, the obtained biases
study, calibration curves were constructed on six sepa- from target values in the commercial normal and
rate occasions by using original applications for PT abnormal control samples ranged from 1.1% to 8.6%,
and Fib calibrations. Criteria for the acceptance of and from 2.3% to 18.2%, respectively.
obtained CVs were <5% for PT, APTT (NCCLS, 1996)
and PC (Woods, Kitchen & Preston, 1999; Meijer,
Precision of calibration curves
Haverkate & Kluft, 2006), <7% for Fib (NCCLS,
2001), and <10% for factor assays (NCCLS, 1997). For The results of precision of calibration curves are
the method comparison study, results obtained on the shown in Table 4. The obtained CVs for all dilution
ACL TOP were compared with those obtained on one
of Dade Behring analyzers: Behring Coagulation Sys-
tem (BCS) and Behring Coagulation Timer (BCT), by Table 2. Within run precision in normal and abnormal
analyzing plasma samples (n = 44–241) with different plasma samples
values over the whole concentration range. Results of Normal Abnormal
FXIII : Ag obtained on the ACL TOP were compared plasma sample plasma sample
with FXIII activities (FXIII : Act) from BCS. The
CV CV
throughput study was performed by using 100 routine
Test n Mean ± SD (%) Mean ± SD (%)
plasma samples with PT/APTT/Fib requests and pro-
cessed on the ACL TOP and BCS. PT
% 20 114.6 ± 2.4 2.1 61.8 ± 1.0 1.6
INR 20 0.94 ± 0.01 0.9 1.28 ± 0.01 1.0
Statistical analysis APTT1 20 31.6 ± 0.5 1.6 50.6 ± 1.0 2.0
(SynthASil), s
The MedCalc program, version 4.10 for Windows, was APTT2 20 28.6 ± 0.4 1.6 56.1 ± 1.3 2.4
used for statistical analysis. In the case of normal dis- (APTT-SP), s
tribution, data were expressed as mean values with Fib, g/l 20 2.9 ± 0.1 3.5 5.6 ± 0.4 7.7
SD, and correlation co-efficient and Student t-test PC, % 10 112.3 ± 1.9 1.7 55.5 ± 2.9 5.2
were performed. When data were not normally dis- FVII, % 10 108.2 ± 3.9 3.6 34.4 ± 1.1 3.1
FVIII, % 10 84.2 ± 4.6 5.5 33.6 ± 1.2 3.5
tributed, median values with ranges, Wilcoxon’s
FXIII : Ag, % 10 113.6 ± 5.2 4.6 68.1 ± 2.0 2.9
matched pairs t-test, and the Spearman’s co-efficient
Ó 2007 The Authors
Journal compilation Ó 2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2009, 31, 26–35
30 M. MILOS ET AL. EVALUATION OF ACL TOP – HEMOSIL REAGENT SYSTEM

Table 3. Between run precision and accuracy in commercial normal and abnormal plasma samples in different runs
over a period of 10 days

HemosIL Normal Control (lot no. HemosIL Low Abnormal Special Test Level 2
N0957778) Control (lot no. N0957718) (lot no. N0957718)

Bias from Bias from Bias from

target CV target CV target
Test Mean ± SD CV (%) value (%) Mean ± SD (%) value (%) Mean ± SD (%) value (%)

PT
% 104.9 ± 8.1 7.7 6.0 30.1 ± 0.01 3.3 3.3 – – –
INR 0.98 ± 0.03 3.0 3.0 2.01 ± 0.1 3.1 )2.9 – – –
APTT1 (SynthASil), s 28.0 ± 0.9 3.3 )1.1 47.4 ± 1.3 2.7 )2.3 – – –
APTT2 (APTT-SP), s 31.3 ± 1.2 3.9 3.0 52.3 ± 2.4 4.6 8.3 – – –
Fib, g/l 3.3 ± 0.1 3.4 5.1 2.6 ± 0.2 6.5 5.7 – – –
PC, % 92.9 ± 1.9 2.0 )5.2 – – – 26.0 ± 3.8 14.8 18.2
FVII, % 88.6 ± 4.8 5.4 )8.6 – – – 24.0 ± 1.9 7.9 4.3
FVIII, % 87.9 ± 5.2 5.9 )6.4 – – – 29.3 ± 4.0 13.7 8.5
FXIII : Ag, % 73.3 ± 3.2 4.3 )3.6 – – – 31.8 ± 1.7 5.3 2.6

The agreement between results obtained on differ-

Table 4. Precision of the calibration curves for PT and
Fib ent analyzers is demonstrated in difference-plots,
according to Bland and Altman, which show differ-
Calibration ences between results over the whole concentration
plasma (%) PT (%) Fib (g/l) Mean ± SD (s) CV (%) range (Figure 1). Despite the good correlation co-effi-
cient for PT% and low mean difference of 0.8%, there
PT 100 100.0 – 12.4 ± 0.2 1.3
50 50.0 – 19.6 ± 0.3 1.6 was an important dispersion of differences (±1.96
25 25.0 – 35.1 ± 0.8 2.3 SD = 24.9%) in PT% testing (Figure 1a). The overall
Fib 150 – 4.35 12.6 ± 0.6 4.8 mean difference for PT INR was 0.16. For higher PT
100 – 2.90 16.7 ± 0.8 5.0 INR values obtained on BCS, there was a trend
60 – 1.74 23.9 ± 1.2 5.0
towards lower PT INR values on the ACL TOP
40 – 1.16 33.0 ± 2.1 6.3
30 – 0.87 41.2 ± 2.9 7.0 (Figure 1b). In APTT testing (Figure 1c,d), similar
mean differences were found for APTT1 (0.10) and
APTT2 ()0.08). Mean difference for Fib was )0.88 g/l,
with wide dispersion of differences (±1.96SD = 1.5 g/
points of PT and Fib calibration curves were within l), and with 86% higher results on the ACL TOP than
the criteria for acceptance. Higher, but uniform CVs on BCS (Figure 1e). Mean differences for FVII
were obtained for the precision of Fib calibration and FVIII were )4.6% and 3.4%, respectively
curve, with increasing tendency in higher dilutions of (Figure 1g,h). For FVII, nonhomogeneous distribution
calibration plasma. of differences was shown, with higher negative devia-
tions for higher levels of FVII on the BCT.
The results obtained in comparison study according
Method comparison study
to reference intervals for the ACL TOP (reference
The results of correlation study between the ACL intervals proposed by the manufacturer, except for
TOP and BCS/BCT are shown in Table 5. The PT%) and for the BCS/BCT (laboratory specific refer-
obtained correlation co-efficients were between ence intervals) are presented in Table 6. The number
0.709 and 0.955. Statistically significant difference in of discrepant results within and outside the reference
results was obtained for PT INR, APTT1, APTT2, Fib, intervals on the BCS/BCT was from 9.4% to 56.8%
and PC. and from 5.4% to 50%, respectively. A total number
Ó 2007 The Authors
Journal compilation Ó 2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2009, 31, 26–35
 M. MILOS ET AL. EVALUATION OF ACL TOP – HEMOSIL REAGENT SYSTEM 31

Table 5. Results of correlation study between the ACL TOP and BCS/BCT analyzers

Test n ACL TOP BCS/BCT P-value r

PT
% 241 80.0 (8.0–143.8) 87.0 (6.0–131.0) 0.247 0.913
INR 93 1.88 (0.94–5.41) 1.63 (0.91–6.65) 0.037* 0.899
APTT1 (SynthASil), ratio 208 0.95 (0.63–3.87) 0.98 (0.74–4.28) <0.001* 0.730
APTT2 (APTT-SP), ratio 109 1.14 (0.73–4.01) <0.001* 0.709
Fib, g/l 201 4.6 (0.4–11.2) 3.5 (0.4–10.6) <0.001* 0.923
PC, % 59 87.6 (18.8–186.9) 91.9 (15.0–173.8) 0.002* 0.955
FVII, % 44 87.8 ± 32.0 83.2 ± 29.8 0.150 0.775
FVIII, % 80 73.1 (3.1–253.4) 83.5 (1.0–303.0) 0.516 0.944
FXIII : Ag, % 66 89.5 ± 36.7 – 0.153 0.928
FXIII : Act, % 66 – 92.0 ± 34.8

In the case of normal distribution – data expressed as mean values with SD; P-values obtained using Student t-test for
paired samples; r, correlation co-efficient. In the case of non-normal distribution – data expressed as median values
with ranges; P-values obtained using Wilcoxon’s matched pairs T-test; r, Spearman’s co-efficient of rank correlation.
*Statistically significant.

of discrepant results ranged from 9.7% to 42.4%.
When analyzing APTT, a total of 33.7% of discrepant
results for APTT1 and 30.3% for APTT2 were
obtained, with a similar number of discrepant results
within and outside the reference intervals. The high-
est number of discrepant results was obtained for PC
and FVII (56.8% and 50%, respectively), although the
mean differences were not high (4% and )4.6%,
respectively; Figure 1). FXIII : Ag results obtained on
the ACL TOP were correlated to FXIII : Act from the
BCS. Results and regression line (Y = 0.98X ) 0.34)
are presented in Figure 2.
Throughput study
The results of throughput study are presented in
Table 7. The total time needed to complete all analy-
ses on the ACL TOP and on the BCS was 3 h 20 min
and 2 h 9 min, and the calculated number of tests
performed per hour was 151 and 212, respectively.
DISCUSSION
Nowadays coagulation laboratories are faced with an
increasing number of requests for different coagula-
tion assays in daily laboratory practice. Automation in
coagulation testing has enabled laboratories to func-
tion more cost-effectively by improving the through-
Ó 2007 The Authors
Journal compilation Ó 2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2009, 31, 26–35
put of samples and by reliability of coagulation assays
(Monagle et al., 2006). Currently, there are a number
of different automated coagulation analyzers that are
commercially available and designed for testing a
large number of samples by combining different
technologies.
In the present study, the coagulation testing system
composed of the ACL TOP analyzer and HemosIL
reagent group was evaluated. The evaluation included
clotting (PT, APTT, Fib, FVII, and FVIII), chromogenic
(PC) and immunological (FXIII : Ag) assays. To our
knowledge, it is the first evaluation of the coagulation
testing system, which included the ACL TOP analyzer
and HemosIL reagents, calibrators and controls.
Concerning within- and between-run precision,
the obtained CVs were within the criteria for accep-
tance in most evaluated assays, and results are similar
to already published evaluations of other fully-auto-
mated coagulation analyzers with similar characteris-
tics, such as the STAR (Flanders et al., 2002), and CA
7000 (Fischer et al., 2006; Dorn-Beineke et al., 2005),
as well as of the ACL TOP (Appert-Flory et al., 2007;
Eschwège, Catillon & Robert, 2006).
As there is little information available concerning
accuracy of measurements within the field of hemo-
stasis, the obtained results are difficult to interpret.
However, PT INR, APTT1, Fib, and PC fulfilled the
desirable or minimum goals for biases given by Ricós
32 M. MILOS ET AL. EVALUATION OF ACL TOP – HEMOSIL REAGENT SYSTEM

(a) 40 (b) 3.0

30 2.5
+1.96 SD

PT INR BCS - ACL TOP

PT (%) BCS - ACL TOP
20 24.1 2.0
10
1.5
Mean
0 +1.96 SD
- 0.8 1.0
–10 1.04
0.5
–20 Mean
–1.96 SD
–30 0.0 0.16
–25.7
–40 –0.5 –1.96 SD
–50 –1.0 –0.73
0 20 40 60 80 100 120 140 0 1 2 3 4 5 6 7
AVERAGE of PT (%) BCS and ACL TOP AVERAGE of PT INR BCS and ACL TOP

(c) 3.0 (d) 1.5

APTT1 (ratio) BCS - ACL TOP

APTT2 (ratio) BCS - ACL TOP

2.5
1.0
2.0

0.5 +1.96 SD
1.5
0.43
1.0 Mean
0.0
+1.96 SD
0.63 –0.08
0.5
Mean 0.10 –0.5 –1.96 SD
0.0
–1.96 SD –0.58
–0.44
–0.5 –1.0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
AVERAGE of APTT1(ratio) BCS and ACL TOP AVERAGE of APTT2 (ratio) BCS and ACL TOP

(e) 1.5 (f) 25

+1.96 SD
1.0 20
+1.96 SD 20.9
Fib (g/L) BCS - ACL TOP

PC (%) BCS - ACL TOP

0.5 15
0.62
0.0 10
–0.5 5 Mean
Mean 4.0
–1.0 0
–0.88
–1.5 –5
–2.0 –10 –1.96 SD
–1.96 SD
–2.5 –15 –13.0
–2.39
–3.0 –20
0 2 4 6 8 10 12 0 20 40 60 80 100 120 140 160 180 200
AVERAGE of Fib (g/L) BCS and ACL TOP AVERAGE of PC (%) BCS and ACL TOP

(g) 100 (h) 100

80 +1.96 SD
FVII (%) BCS - ACL TOP

FVIII (%) BCS - ACL TOP

60 50
55.7
40 +1.96 SD
Mean
36.3 0
20 3.4
0 Mean –1.96 SD
–50
–20 –4.6 –48.9

–40 –1.96 SD –100

–60 –45.5

–80 –150
20 40 60 80 100 120 140 160 180 0 50 100 150 200 250 300
AVERAGE of FVII (%) BCS and ACL TOP AVERAGE of FVIII (%) BCS and ACL TOP

Figure 1. Comparison performed according to Bland and Altman of test results obtained on the ACL TOP and BCS/
BCT analyzers. (a) PT%, (b) PT INR, (c) APTT1, (d) APTT2, (e) Fib, (f) PC, (g) FVII and (h) FVIII.

Ó 2007 The Authors

Journal compilation Ó 2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2009, 31, 26–35
 M. MILOS ET AL. EVALUATION OF ACL TOP – HEMOSIL REAGENT SYSTEM 33

Table 6. Discrepancy between test results obtained on the ACL TOP and BCS/BCT analyzers (according to reference
intervals)

No. discrepant test results (n, %) on ACL TOP

Results outside
Reference Reference Results within the the reference
intervals for intervals for reference interval interval on
Test ACL TOP* BCS/BCT† n on BCS/BCT BCS/BCT Total

PT
%† 70.0 70.0 241 17/147 (11.6) 11/94 (11.7) 28/241 (11.6)
INR‡ 2.00–3.50 2.00–3.50 93 5/19 (26.3) 4/74 (5.4) 9/93 (9.7)
APTT1 (SynthASil), ratio 0.89–1.11 0.84–1.16 208 45/134 (33.6) 25/74 (33.8) 70/208 (33.7)
APTT2 (APTT-SP), ratio 0.84–1.16 109 20/62 (32.3) 13/47 (27.7) 33/109 (30.3)
FIB, g/l 2.4–5.0 1.8–4.1 201 18/113 (15.9) 12/88 (13.6) 30/201 (14.9)
PC, % 93.9–158.7 70.0–140.0 59 21/37 (56.8) 4/22 (18.2) 25/59 (42.4)
FVII, % 50.0–129.0 70.0–120.0 44 3/32 (9.4) 6/12 (50) 9/44 (20.4)
FVIII, % 50.0–150.0 60.0–180.0 80 7/41 (17.1) 5/39 (12.8) 12/80 (15.0)

Results are expressed as the number (n) and percentage (%) of discrepant test results within the reference interval on
the BCS/BCT, results outside the reference interval on the BCS/BCT, and the total number of discrepant test results.
*Reference intervals proposed for the ACL TOP by the manufacturer.
†Laboratory specific reference intervals.
‡For PT INR, results were analyzed according to therapeutic range.

220
Table 7. Results of the throughput study (PT/APTT/Fib,
200
n = 100)
180
FXIII:Ag (%) ACL TOP

160
ACL TOP BCS
140
120
Preanalytical time (h:min)* 1:06 0:35
100
Time to first result (h:min) 0:03 0:08
80
Time to last result (h:min) 1:59 1:25
60
Total time (h:min)† 3:20 2:09
40
No. samples performed per hour 50 71
20
20 40 60 80 100 120 140 160 180 200 No. tests performed per hour 151 212
FXIII: Act (%) BCS
*Comprise time to prepare the analyzer, (and reconstitu-
tion and stabilization of) reagents and control samples.
Figure 2. Correlation of the HemosIL FXIII Antigen on
†Comprise preanalytical time, analytical time and time
the ACL TOP with the Berichrom FXIII activity on
for maintenance and shutdown of the analyzers.
the BCS (Y = 0.98X ) 0.34, r = 0.928).

et al. (1999) and Meijer et al. (2006), while minimum
goals for PT%, APTT2, FVII, and FVIII could not be
achieved in at least one control sample. For FXIII,
there is until now no published data concerning bio-
logical variation (Ricós et al., 1999).
Although satisfactory precision of PT and Fib cali-
bration curves was obtained (Table 4), the disadvan-
tage of the proposed PT calibration was the last
Ó 2007 The Authors
Journal compilation Ó 2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2009, 31, 26–35
dilution point of 25%. As fibrinogen concentration in
this dilution point was 0.725 g/l, and as low fibrino-
gen concentration is known to be a limiting factor for
PT measurement, it was not possible to extend the
calibration furthermore, making PT results below 25%
questionable.
Despite good correlation between results obtained
on the ACL TOP and BCS/BCT, statistically significant
34 M. MILOS ET AL. EVALUATION OF ACL TOP – HEMOSIL REAGENT SYSTEM
difference was established for PT INR, APTT1, APTT2,
Fib, and PC. Besides the correlation, which is a mea-
sure of relationship between the test results of two
different methods, the testing of agreement, that
includes numerical identity between the test results of
two different methods, is also necessary to perform as
it provides additional information (Meijer, 2003). The
agreement between results analyzed using difference-
plots (Figure 1) showed that mean differences could
be of clinical relevance, i.e., in case of PT INR this dif-
ference may lead to inappropriate anticoagulant dos-
age decisions.
A high number of discrepant results obtained for
APTT, PC and FVII despite the low mean difference
indicate that the reference intervals proposed by the
manufacturer for these assays on the ACL TOP seem
to be inappropriate, and that new reference intervals
should be determined.
The Scientific Subcommittee on FVIII and FIX of
the Scientific and Standardization Committee of the
International Society on Thrombosis and Haemosta-
sis has recently recommended the classification of
severe, moderate and mild hemophilia A on the
basis of biological FVIII levels of <1, 1–5, and >5%,
respectively, rather than on the clinical severity of
affected individuals (Preston et al., 2004; White et al.,
2001). According to that, coagulation laboratory has
a vital role in the diagnosis and classification of
hemophilia patients, and results of determinations
should be both precise and accurate (Chuansumrit,
McCraw & Preston, 2004). In our study, we
encountered problems with FVIII determination on
the ACL TOP in plasma samples with FVIII<10%.
We observed that the analyzer automatically remea-
sured all samples with FVIII activities below 10%,
using different dilution of plasma and factor diluent
on the same calibration curve, although the stated
linearity and test range for FVIII is 1–150% and
0.8–150%, respectively. However, the analyzer gen-
erated messages with both results, making them
Journal compilation Ó 2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2009, 31, 26–35
uncertain. Moreover, nine of 18 (50%) severe
hemophilia A patients would not be properly classi-
fied (data not shown), which is in concordance with
Barrowcliffe (2004) who stressed that the lower
limit of a one-stage method (1%) may be difficult
to achieve with some reagents.
Although FXIII : Act assay has been the method of
choice, HemosIL FXIII : Ag, as an immunoturbidi-
metric assay for measuring the subunit A of the FXIII
tetramer, offers new opportunities for FXIII measure-
ment. In our study, this assay showed good perfor-
mances in precision, accuracy, as well as correlation
with FXIII : Act. The investigation of agreement could
not be performed because of nonavailability of any
other assay for FXIII : Ag measurement in our labora-
tory.
During evaluation, we observed some potential
advantages and disadvantages of the evaluated sys-
tem. The analyzer enables the real time monitoring
of each sample status on board together with the
possibility of inspection of every single stored reac-
tion curve. Although the software offers the over-
view of reagent volume on board, it is not possible
to see how much reagent is needed to process the
job list. As a consequence, in the case when there
is not enough volume of any component needed for
the reaction, the instrument gives the warning but
proceeds with the procedure and consumes other
components not completing the analysis. The limita-
tion of the software is also that all relevant data for
reagents, controls and calibrators must be entered
manually by the operator.
Although the ACL TOP was found to have a num-
ber of performances similar to other modern auto-
mated coagulation analyzers in regard to its ability to
analyze a high number of samples, the obtained
throughput that was by 29% lower than for the BCS,
together with inappropriate package volumes for some
special tests makes the analyzer suitable rather for
mid-sized laboratories.
Ó 2007 The Authors

REFERENCES
Appert-Flory A., Fischer F., Jambou D. &
Toulon P. (2007) Evaluation and perfor-
mance characteristics of the automated
coagulation analyzer ACL TOP. Throm-
bosis Research 120, 733–743.
Barrowcliffe T.W. (2004) Monitoring hae-
mophilia severity and treatment: new
or old laboratory tests? Haemophilia
10(Suppl. 4), 109–114.
Bland J.M. & Altman D.J. (1986) Statisti-
cal methods for assessing agreement
between two methods of clinical mea-
surement. Lancet 1, 307–310.
Chuansumrit A., McCraw A. & Preston
E.F. (2004) Essential issues of labora-
tory investigation for patients with hae-
mophilia and bleeding disorders.
Haemophilia 10, 105–108.
Clauss A. (1957) Gerinnungsphysiologi-
sche Schnellmethode zur Bestimmung
des Fibrinogens. Acta Haematologica 17,
237–246.
Dorn-Beineke A., Dempfle C.E., Bertsch
T. & Wisser H. (2005) Evaluation of the
automated coagulation analyzer Sysmex
CA-7000. Thrombosis Research 116,
171–179.
Eschwège V., Catillon C. & Robert A.
(2006) Evaluation of the automated
coagulation analyser ACL TOP (Instru-
mentation Laboratory). Annales de Bio-
logie Clinique 64, 259–264.
Fischer F., Appert-Flory A., Jambou D. &
Toulon P. (2006) Evaluation of the
automated coagulation analyzer Sysmex
Ó 2007 The Authors
Journal compilation Ó 2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2009, 31, 26–35
M. MILOS ET AL. EVALUATION OF ACL TOP – HEMOSIL REAGENT SYSTEM 35
CA 7000. Thrombosis Research 117,
721–729.
Flanders M.M., Crist R., Safapour S. &
Rodgers G.M. (2002) Evaluation and
performance characteristics of the STA-
R coagulation analyzer. Clinical Chem-
istry 48, 1622–1624.
Meijer P. (2003) Method comparison: cor-
relation is not enough to decide about
agreement and clinical accuracy.
Thrombosis and Haemostasis 90, 555–
556.
Meijer P., Haverkate F. & Kluft C. (2006)
Performance goals for the laboratory
testing of antithrombin, protein C and
protein S. Thrombosis and Haemostasis
96, 584–589.
Monagle P., Barnes C., Ignjatovic V.,
Furmedge J., Newall F., Chan A., De
Rosa L., Hamilton S., Ragg P., Robinson
S., Auldist A., Crock C., Roy N. & Row-
lands S. (2006) Developmental haemo-
stasis. Thrombosis and Haemostasis 95,
362–372.
NCCLS (1996) One-Stage Prothrombin
Time (PT) Test and Activated Partial
Thromboplastin Time (APTT) Test;
Approved Guideline. NCCLS Document
H47-A (ISBN 1-56238-301-9). NCCLS,
Wayne, PA.
NCCLS (1997) Determination of Factor
Coagulant Activities; Approved Guide-
line. NCCLS Document H48-A (ISBN 1-
56238-321-3). NCCLS, Wayne, PA.
NCCLS (2001) Procedure for the Determi-
nation of Fibrinogen in Plasma;
Approved Guideline-Second Edition.
NCCLS Document H30-A2 (ISBN 1-
56238-439-2). NCCLS, Wayne, PA.
Preston F.E., Kitchen S., Jennings I.,
Woods T.A.L. & Makris M. (2004) SSC/
ISTH classification of hemophilia A: can
hemophilia center laboratories achieve
the new criteria? Journal of Thrombosis
and Haemostasis 2, 271–274.
Ricós C., Alvarez V., Cava F., Garcia-Lario
J.V., Hernandez A., Jimenez C.V., Min-
chinela J., Perich C. & Simon M. (1999)
Current database on biological varia-
tion: pros, cons and progress. Scandina-
vian Journal of Clinical and Laboratory
Investigation 59, 491–500.
Walenga J.M. & Fareed J. (1994) Auto-
mation and quality control in the coag-
ulation laboratory. Clinics in Laboratory
Medicine 14, 709–728.
White G.C., Rosendaal F., Aledort L.M.,
Lusher J.M., Rothschild C. & Ingerslev
J. (2001) Definitions in Hemophilia:
Recommendations of the Scientific Sub-
committee on FVIII and FIX of the SSC
of ISTH. Thrombosis and Haemostasis
85, 560.
Woods T.A.L., Kitchen S. & Preston F.E.
(1999) Quality assessment of haemo-
static assays and external quality assess-
ment schemes. In: Laboratory
Techniques in Thrombosis – A Manual
(eds J. Jespersen, R.M. Bertina & F.
Haverkate), 2nd revised edition of
ECAT assay procedures, pp. 29–36. Klu-
wer Academic Publishers, Dordrecht,
The Netherlands.