Professional Documents
Culture Documents
*
Clinical Institute of Laboratory SUMMARY
Diagnosis, Clinical Hospital Center
Zagreb University School of ACL TOP is a fully automated coagulation analyzer, designed for
Medicine, Zagreb, Croatia, simultaneous measurement of routine and special coagulation param-
†
Faculty of Pharmacy and
Biochemistry, University of Zagreb,
eters. We evaluated analytical and technical performance characteris-
Zagreb, Croatia tics of the coagulation system composed of the ACL TOP analyzer and
HemosIL reagent group for the determination of routine clotting (PT,
Correspondence: APTT, fibrinogen, FVII, and FVIII), chromogenic (protein C) and
Dr Renata Zadro, Clinical Institute
of Laboratory Diagnosis, Zagreb immunological assays (FXIII antigen). Within run and between run
Clinical Hospital Center, Kispaticeva CVs ranged from 0.9% to 7.7% and from 2.0% to 14.8% respectively.
12, 10000 Zagreb, Croatia. The obtained CVs for imprecision of calibration curves were <5% for PT
Tel.: +385 1 2388 247;
and <7% for fibrinogen. The method comparison study showed good
Fax: +385 1 2312 079;
E-mail: rzadro@mef.hr correlation between results obtained on the ACL TOP and BCS/BCT
analyzers, with correlation co-efficients ranging from 0.709 to 0.955,
doi:10.1111/j.1751-553X.2007.00999.x but with significantly different results for PT INR, APTT, fibrinogen and
Received 2 May 2007; accepted for
protein C, and wide dispersion of differences observed in difference
publication 2 October 2007 plots for most assays. Despite good correlation and agreement for FVIII,
problems in measuring FVIII<10% were encountered. The effective
Keywords throughput for the ACL TOP and BCS was 151 and 212 PT/APTT/
Coagulation analyzer, ACL TOP, fibrinogen tests per hour, respectively. Although the ACL TOP
HemosIL reagents, evaluation
is designed to run multiple assays on a large number of samples,
software limitations make the instrument suitable rather for mid-sized
laboratories.
testing system composed of the ACL TOP coagulation for immunologic measurements. The control module
analyzer and HemosIL reagent group (Instrumentation consists of the software running under Windows 2000
Laboratory, Lexington, MA, USA) under routine labo- Operating System (version 2.1) (Microsoft Corpora-
ratory conditions. We tested prothrombin time (PT), tion, Redmond, WA, USA) and touch-screen. The
activated partial thromboplastin time (APTT), fibrino- analyzer has the software capabilities for automatic
gen (Fib), factor VII (FVII), factor VIII (FVIII) and pro- rerun, automatic reflex testing, STAT sample loading,
tein C (PC) activities, and FXIII antigen (FXIII : Ag). and factor parallelism testing. There is a large results
The evaluation addressed several issues, including database (20 000 samples with up to 30 assays per
available methodologies, validation of performance, sample) and a possibility of access to reaction curves.
throughput study, reagent and patient sample on- Daily maintenance is a semiautomatic procedure and
board capabilities and ease of operation. usually lasts approximately 7 min.
Table 1. Reagents and calibrators used on the ACL TOP, BCS and BCT
chloride (CaCl2) solution was added (0.025 mol/l after adding of the 0.02 mol/l CaCl2 solution. The
CaCl2 for APTT-SP and 0.02 mol/l CaCl2 for analyzer automatically remeasured all samples with
SynthASil). FVIII activities <10%, using different dilution of
Fib was measured according to Clauss (1957) by plasma and factor diluent on the same calibration
adding plasma sample, diluted 1 : 10 in Factor Diluent curve.
solution, to thrombin solution (35 UNIH/ml) (Fibrino- For FXIII : Ag determination, plasma sample,
gen C-XL). Samples with Fib concentration >6.5 and diluted 1 : 10 in FXIII Diluent solution, was preincu-
<0.8 g/l were remeasured using Fib-high and Fib-low bated at 37 °C. After adding of FXIII Buffer, and incu-
assay procedures, respectively. bation, FXIII Latex reagent (latex polystyrene particles
PC activity was determined using a chromogenic coated with a rabbit polyclonal antibody against the
assay by incubating plasma sample, diluted 1 : 4 in A-subunit of FXIII) was added and the change in
Protein C Diluent solution, and PC activator solution OD671 nm was recorded.
(fraction from the venom of Agkistrodon contortrix con- Measured clotting times of coagulation methods
tortrix, 0.4 U/ml) for 270 s. After adding PC Chromo- (PT, Fib, FVII, and FVIII) or optical densities in chro-
genic substrate S2366 (1.5 mg/ml), the change in mogenic (PC) or immunological (FXIII : Ag) methods
OD405 nm was recorded. were automatically converted into final results of spe-
For FVII activity measurement, FVII deficient cific activities or antigen concentrations by using the
plasma was added to plasma sample, diluted 1 : 10 in appropriate calibration curve. International normal-
Factor Diluent solution. After incubation at 37 °C, ized ratio (INR) was calculated by the analyzer using
final measurement was performed after the addition international sensitivity index value (ISI = 0.800) sta-
of RecombiPlasTin. ted for the ACL TOP by the manufacturer. An auto-
For FVIII activity measurement, FVIII deficient matic rerun test with different sample dilution,
plasma was mixed with plasma sample, diluted 1 : 10 preprogrammed by the manufacturer, was used in
in Factor Diluent solution, and SynthASil. After incu- cases of levels above or below the measuring limits of
bation at 37 °C, final measurement was performed calibration curves.
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 29
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)
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
Table 5. Results of correlation study between the ACL TOP and BCS/BCT analyzers
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%. put of samples and by reliability of coagulation assays
When analyzing APTT, a total of 33.7% of discrepant (Monagle et al., 2006). Currently, there are a number
results for APTT1 and 30.3% for APTT2 were of different automated coagulation analyzers that are
obtained, with a similar number of discrepant results commercially available and designed for testing a
within and outside the reference intervals. The high- large number of samples by combining different
est number of discrepant results was obtained for PC technologies.
and FVII (56.8% and 50%, respectively), although the In the present study, the coagulation testing system
mean differences were not high (4% and )4.6%, composed of the ACL TOP analyzer and HemosIL
respectively; Figure 1). FXIII : Ag results obtained on reagent group was evaluated. The evaluation included
the ACL TOP were correlated to FXIII : Act from the clotting (PT, APTT, Fib, FVII, and FVIII), chromogenic
BCS. Results and regression line (Y = 0.98X ) 0.34) (PC) and immunological (FXIII : Ag) assays. To our
are presented in Figure 2. knowledge, it is the first evaluation of the coagulation
testing system, which included the ACL TOP analyzer
and HemosIL reagents, calibrators and controls.
Throughput study
Concerning within- and between-run precision,
The results of throughput study are presented in the obtained CVs were within the criteria for accep-
Table 7. The total time needed to complete all analy- tance in most evaluated assays, and results are similar
ses on the ACL TOP and on the BCS was 3 h 20 min to already published evaluations of other fully-auto-
and 2 h 9 min, and the calculated number of tests mated coagulation analyzers with similar characteris-
performed per hour was 151 and 212, respectively. 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;
DISCUSSION
Eschwège, Catillon & Robert, 2006).
Nowadays coagulation laboratories are faced with an As there is little information available concerning
increasing number of requests for different coagula- accuracy of measurements within the field of hemo-
tion assays in daily laboratory practice. Automation in stasis, the obtained results are difficult to interpret.
coagulation testing has enabled laboratories to func- However, PT INR, APTT1, Fib, and PC fulfilled the
tion more cost-effectively by improving the through- desirable or minimum goals for biases given by Ricós
Ó 2007 The Authors
Journal compilation Ó 2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2009, 31, 26–35
32 M. MILOS ET AL. EVALUATION OF ACL TOP – HEMOSIL REAGENT SYSTEM
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
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
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
–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.
Table 6. Discrepancy between test results obtained on the ACL TOP and BCS/BCT analyzers (according to reference
intervals)
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 dilution point of 25%. As fibrinogen concentration in
goals for PT%, APTT2, FVII, and FVIII could not be this dilution point was 0.725 g/l, and as low fibrino-
achieved in at least one control sample. For FXIII, gen concentration is known to be a limiting factor for
there is until now no published data concerning bio- PT measurement, it was not possible to extend the
logical variation (Ricós et al., 1999). calibration furthermore, making PT results below 25%
Although satisfactory precision of PT and Fib cali- questionable.
bration curves was obtained (Table 4), the disadvan- Despite good correlation between results obtained
tage of the proposed PT calibration was the last on the ACL TOP and BCS/BCT, statistically significant
Ó 2007 The Authors
Journal compilation Ó 2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2009, 31, 26–35
34 M. MILOS ET AL. EVALUATION OF ACL TOP – HEMOSIL REAGENT SYSTEM
difference was established for PT INR, APTT1, APTT2, uncertain. Moreover, nine of 18 (50%) severe
Fib, and PC. Besides the correlation, which is a mea- hemophilia A patients would not be properly classi-
sure of relationship between the test results of two fied (data not shown), which is in concordance with
different methods, the testing of agreement, that Barrowcliffe (2004) who stressed that the lower
includes numerical identity between the test results of limit of a one-stage method (1%) may be difficult
two different methods, is also necessary to perform as to achieve with some reagents.
it provides additional information (Meijer, 2003). The Although FXIII : Act assay has been the method of
agreement between results analyzed using difference- choice, HemosIL FXIII : Ag, as an immunoturbidi-
plots (Figure 1) showed that mean differences could metric assay for measuring the subunit A of the FXIII
be of clinical relevance, i.e., in case of PT INR this dif- tetramer, offers new opportunities for FXIII measure-
ference may lead to inappropriate anticoagulant dos- ment. In our study, this assay showed good perfor-
age decisions. mances in precision, accuracy, as well as correlation
A high number of discrepant results obtained for with FXIII : Act. The investigation of agreement could
APTT, PC and FVII despite the low mean difference not be performed because of nonavailability of any
indicate that the reference intervals proposed by the other assay for FXIII : Ag measurement in our labora-
manufacturer for these assays on the ACL TOP seem tory.
to be inappropriate, and that new reference intervals During evaluation, we observed some potential
should be determined. advantages and disadvantages of the evaluated sys-
The Scientific Subcommittee on FVIII and FIX of tem. The analyzer enables the real time monitoring
the Scientific and Standardization Committee of the of each sample status on board together with the
International Society on Thrombosis and Haemosta- possibility of inspection of every single stored reac-
sis has recently recommended the classification of tion curve. Although the software offers the over-
severe, moderate and mild hemophilia A on the view of reagent volume on board, it is not possible
basis of biological FVIII levels of <1, 1–5, and >5%, to see how much reagent is needed to process the
respectively, rather than on the clinical severity of job list. As a consequence, in the case when there
affected individuals (Preston et al., 2004; White et al., is not enough volume of any component needed for
2001). According to that, coagulation laboratory has the reaction, the instrument gives the warning but
a vital role in the diagnosis and classification of proceeds with the procedure and consumes other
hemophilia patients, and results of determinations components not completing the analysis. The limita-
should be both precise and accurate (Chuansumrit, tion of the software is also that all relevant data for
McCraw & Preston, 2004). In our study, we reagents, controls and calibrators must be entered
encountered problems with FVIII determination on manually by the operator.
the ACL TOP in plasma samples with FVIII<10%. Although the ACL TOP was found to have a num-
We observed that the analyzer automatically remea- ber of performances similar to other modern auto-
sured all samples with FVIII activities below 10%, mated coagulation analyzers in regard to its ability to
using different dilution of plasma and factor diluent analyze a high number of samples, the obtained
on the same calibration curve, although the stated throughput that was by 29% lower than for the BCS,
linearity and test range for FVIII is 1–150% and together with inappropriate package volumes for some
0.8–150%, respectively. However, the analyzer gen- special tests makes the analyzer suitable rather for
erated messages with both results, making them mid-sized laboratories.