Thrombosis Research 133 (2014) 927–935

Contents lists available at ScienceDirect

Thrombosis Research
journal homepage: www.elsevier.com/locate/thromres

Regular Article

Evaluation and performance characteristics of the Q Hemostasis Analyzer,

an automated coagulation analyzer
Pierre Toulon a,b,⁎, Florence Fischer b, Anny Appert-Flory b, Didier Jambou b
a
Université Nice Sophia-Antipolis, Faculté de Médecine, Département d’Hématologie, Nice, France
b
CHU, Hôpital Pasteur, Service d’Hématologie Biologique, Nice, France

a r t i c l e i n f o a b s t r a c t

Article history: The Q Hemostasis Analyzer (Grifols, Barcelona, Spain) is a fully-automated random-access multiparameter analyz-
Received 15 January 2014 er, designed to perform coagulation, chromogenic and immunologic assays. It is equipped with a cap-piercing
Received in revised form 21 February 2014 system. The instrument was evaluated in a hemostasis laboratory of a University Hospital with respect to its tech-
Accepted 21 February 2014
nical features in the determination of coagulation i.e. prothrombin time (PT), activated partial thromboplastin
Available online 28 February 2014
time (aPTT), thrombin time, fibrinogen and single coagulation factors V (FV) and VIII (FVIII), chromogenic [anti-
Keywords:
thrombin (AT) and protein C activity] and immunologic assays [von Willebrand factor antigen (vWF:Ag) concen-
Coagulation analyzer tration], using reagents from the analyzer manufacturer. Total precision (evaluated as the coefficient of variation)
Evaluation was below 6% for most parameters both in normal and in pathological ranges, except for FV, FVIII, AT and vWF:Ag
Q Hemostasis Analyzer both in the normal and pathological samples. No carryover was detected in alternating aPTT measurement in a
pool of normal plasma samples and in the same pool spiked with unfractionated heparin (N 1.5 IU/mL). The effec-
tive throughput was 154 PT, 66 PT/aPTT, 42 PT/aPTT/fibrinogen, and 38 PT/aPTT/AT per hour, leading to 154 to
114 tests performed per hour, depending of the tested panel. Test results obtained on the Q Hemostasis Analyzer
were well correlated with those obtained on the ACL TOP analyzer (Instrumentation Laboratory), with r between
0.862 and 0.989. In conclusion, routine coagulation testing can be performed on the Q Hemostasis Analyzer with
satisfactory precision and the same apply to more specialized and specific tests.
© 2014 Elsevier Ltd. All rights reserved.

Introduction In that respect, there is a need for independent external evaluations of

The increasing demand for coagulation tests as well as the econom-
ical pressure for reducing staff costs and decreasing reagents budgets
has raised interest in hemostasis laboratory automation [1]. The current
generation of coagulation analyzers is fully automated. Their capabilities
include primary tube sampling, associated in some cases with cap
piercing, dilution capabilities, automatic rerun, and reflex testing. They
can perform basic coagulation tests such as the prothrombin time (PT)
or the activated partial thromboplastin time (aPTT) as well as more so-
phisticated coagulation, chromogenic and immunologic assays [2,3]
using smaller sample and reagent volumes than the manual methods.
As many of such coagulation analyzers are available today, the selection
of the one that meets the local needs is a critical step for each laboratory.
such analyzers, as these data could be of great help. The aim of the
present study was to evaluate the performance characteristics of the Q
Hemostasis Analyzer, a new automated analyzer in the routine practice.
The analyzer was evaluated in a coagulation laboratory of a University
Hospital according to current recommendations [4,5]. The evaluation
addressed several topics including ease of operation, method availabili-
ty, reagent and patient sample on-board capabilities, ability to perform
automatic dilution for calibration, rerun and reflex testing, with a partic-
ular emphasis on the validation of performances.
Materials and Methods
Description of the Q Hemostasis Analyzer

Abbreviations: aPTT, activated partial thromboplastin time; AT, antithrombin; CV, co-
efficient of variation; FV, (coagulation) factor V; FVIII, (coagulation) factor VIII; INR, inter-
national normalized ratio; ISI, international sensitivity index; MNPT, mean normal
prothrombin time; OD, optical density; PC, protein C; PT, prothrombin time; QC, quality
control; SD, standard deviation; TT, thrombin time; UFH, unfractionated heparin; vWF:
Ag, von Willebrand factor antigen.
⁎ Corresponding author at: CHU de Nice, Hôpital Pasteur Service d'Hématologie
Biologique 30, avenue de la Voie Romaine CS 51069 F-06001 Nice Cedex 1, France.
Tel.: +33 4 92 03 87 09; fax: +33 4 92 03 85 95.
E-mail address: toulon.p@chu-nice.fr (P. Toulon).
The Q Hemostasis Analyzer (Grifols SA, Barcelona, Spain) is a fully au-
tomated stand-alone random-access multiparameter benchtop coagu-
lation analyzer, with cap-piercing capabilities. Clotting, chromogenic
and immunoturbidimetric assays can be performed in a true random
fashion. The clot/reaction detection is performed using a photo-optical
unit consisting of eight independent measuring channels that allow
continuous monitoring of reactions at 405 nm with a high resolution
video technology. The cuvette loading area can be filled, even while

http://dx.doi.org/10.1016/j.thromres.2014.02.021
0049-3848/© 2014 Elsevier Ltd. All rights reserved.
928 P. Toulon et al. / Thrombosis Research 133 (2014) 927–935

Table 1
Reagents and calibrators used on the Q Hemostasis Analyzer and the ACL TOP for the comparison study. All reagents were from the analyzer manufacturers, except otherwise stated.

Test Q Hemostasis Analyzer (Grifols) ACL TOP (Instrumentation Laboratory) Methodology

PT DG-PT and DG-Ref HemosIL RecombiPlasTin 2G (IL) Coagulation

aPTT DG-APTT Synth Triniclot APTT HS (TCoag) Coagulation
Fibrinogen DG-FIB L Human and DG-Ref HemosIL Fibrinogen C (IL) and HemosIL Calibration Plasma (IL) Coagulation
Thrombin time DG-TT L Human HemosIL Thrombin Time (IL) Coagulation
Factor V DG-FV + DG-PT and DG-Ref HemosIL Factor V Deficient Plasma (IL) + HemosIL RecombiPlasTin 2G (IL) and Coagulation
Standard Human Plasma ORKL (Siemens)
Factor VIII DG-FVIII + DG-APTT and DG-Ref HemosIL Factor VIII Deficient Plasma (IL) + Triniclot APTT HS (TCoag) and Coagulation
Standard Human Plasma ORKL (Siemens)
Antithrombin DG-Chrom AT and DG-Ref HemosIL Liquid Antithrombin (IL) and Standard Human Plasma ORKL (Siemens) Chromogenic
Protein C DG-Chrom PC and DG-Ref HemosIL Protein C (IL) and Standard Human Plasma ORKL (Siemens) Chromogenic
vWF:Ag DG-Latex VWF:Ag and DG-Ref HemosIL Von Willebrand Factor Antigen (IL) and VWF Calibrator (IL) Immuno-turbidimetry

running, with 2 vertical racks containing 200 unbreakable cuvettes normal routine coagulation profile, patients with single coagulation
each. Up to 30 reagents can be stored on board at 15 °C, with 4 positions factor or inhibitor deficiency, patients with lupus anticoagulant, pa-
under constant stirring. A positive bar-code identification of vials allows tients with liver failure or disseminated intravascular coagulation, and
monitoring of reagent name, batch number, expiration date, ISI (if appli- patients treated with unfractionated heparin (UFH) or vitamin K antag-
cable), stability and available volume, when reagents from the analyzer onist (n variable, depending of the tested parameter). In addition, plas-
manufacturer are used. However, reagents from different manufac- ma samples with optical abnormalities i.e. hemolyzed (n = 18), icteric
turers other than Grifols can be used in their original vials without any (n = 15) and lipemic samples (n = 3) were specifically investigated.
adaptor, thanks to a simple gripping device. In addition to dispense Venous blood was collected into polymer evacuated tubes (Vacuette,
reagents, the revolving multitasking reagent arm pushes cuvettes to Greiner BioOne, Les Ulis, France), containing 0.109 M sodium citrate
the reading area. It is equipped with a pre-heated probe with capacitive (1 vol./9 vol.) according to international recommendations [6]. Plasma
detection allowing “real time” reagent volume monitoring. Samples, was obtained by centrifugation at 2000 x g and +15 °C for 15 minutes.
placed in individual sample holders (Q Sample Holders), can be continu- After having undergone a second centrifugation, the remaining plasma
ously inserted in the analyzer with the use of an original magnetic desk was stored frozen in aliquots at −80 °C until evaluated [7,8].
at ambient temperature, which also enables their release after tests
were completed. Each Q Sample Holder is individually identified using Evaluation Procedure
radio-frequency ID, the TAG RFID reader being located on the front
part of the analyzer. There is a true positive identification of patients’ Precision Testing
samples using an integrated laser bar-code reader. The sample arm is System precision for coagulation i.e. PT, aPTT, fibrinogen, thrombin
equipped with cap-piercing capabilities, allowing the use of primary time (TT), coagulation factors V (FV), and VIII (FVIII), chromogenic i.e.
caped tubes, but decaped tubes and cups can be used at once using antithrombin (AT), and protein C (PC), and immunologic assays [von
the same sample holders without using any adaptor. This arm is Willebrand factor antigen (vWF:Ag)] was assessed according to the
equipped with a probe with capacitive detection allowing sample vol- CLSI guidelines. Preliminary (within-run) precision was evaluated, for
ume monitoring, and a needle equipped with a clot detection system. each test, by repeated testing of a normal and an abnormal frozen plas-
Two wash solutions are required to prime and clean the probe internally ma sample 20 times in the same run (CLSI 10-A3) [9]. Total precision
(Q Autoclean A) and externally (Q Autoclean B), the latter being also used was evaluated by performing 2 runs on 10 separate days, each run
for the shut down maintenance. The user interface consists of a large- consisted of 2 replicates of the same two frozen plasma samples
scaled touch-screen linked to an articulated arm and a software running (normal and abnormal) i.e. 40 determinations (CLSI EP05-A2) [10].
under Windows XP®. The software version available at the time of the The within-run and total precisions were expressed as the coefficient
evaluation (v.3.02) provides a test menu that includes all the tests vali- of variations (CV%) calculated as the standard deviation divided by the
dated by the instrument manufacturer for its own reagents (n = 42 at mean value, for each of the studied parameters.
the time of the experiment, but additional new tests are expected to be
regularly included), which cannot be modified. As it is an open system,
Linearity Analysis
the user has the possibility to adapt its own reagents/methodologies,
When applicable, the linearity of the calibration curves was deter-
using a user-friendly screen. The analyzer has the software capabilities
mined by using serial dilutions of a pool of plasma samples in order to
for automatic rerun, and reflex testing before releasing the sample and
is able to run factor assays at various dilutions simultaneously. The soft-
Table 2
ware also has a complete Quality Control (QC) module able to integrate Preliminary (Intra-assay) precision of the Q Hemostasis Analyzer for prothrombin time
the ranges defined by the manufacturer, customized dynamic control (PT), activated partial thromboplastin time (aPTT), fibrinogen, thrombin time, factor V
ranges or both. It has unlimited QC data storage capacity and includes and VIII, antithrombin activity, protein C activity, and von Willebrand factor antigen
the options of QC planning, Levey-Jennings graph, Westgard rules and concentration (vWF:Ag). The same frozen normal and abnormal plasma samples were
evaluated 20 times simultaneously (in the same run). The results are given as the mean
configurable quality control policies. The LIS interface is based on the
values and standard deviation, with the coefficient of variation (in %) in brackets.
bi-directional host-query-ASTM protocol (not tested in the present eval-
uation). Finally, the maintenance is limited to a 15-minutes daily auto- Normal sample (n = 20) Pathological sample (n = 20)
matic procedure that does not require the presence of a technician. PT (sec) 13.1 ± 0.1 (1.1%) 51.9 ± 0.7 (1.3%)
aPTT (sec) 28.4 ± 0.8 (3.0%) 54.5 ± 2.3 (4.2%)
Fibrinogen (g/L) 3.92 ± 0.09 (2.2%) 1.63 ± 0.06 (3.5%)
Evaluated Samples
Thrombin time (sec) 19.1 ± 0.2 (1.3%) 33.3 ± 1.0 (3.0%)
Factor V (%) 97.4 ± 3.4 (3.5%) 57.2 ± 2.5 (4.4%)
Commercially available normal (DG-C1) and pathological (DG-C2) Factor VIII (%) 175.5 ± 12.1 (6.9%) 47.0 ± 5.3 (11.1%)
lyophilized plasma samples were from Diagnostic Grifols. In addition, Antithrombin (%) 92.9 ± 2.4 (2.6%) 63.3 ± 2.7 (4.3 %)
plasma samples referred to the laboratory for routine coagulation test- Protein C (%) 106.2 ± 1.0 (0.9%) 38.9 ± 1.2 (3.2%)
vWF:Ag (%) 179.2 ± 7.2 (4.0%) 17.8 ± 2.0 (4.4%)
ing were also evaluated. This included plasmas from patients with
 P. Toulon et al. / Thrombosis Research 133 (2014) 927–935 929

Table 3 were investigated: PT, PT + aPTT, PT + aPTT + fibrinogen, and PT +

Total precision of the Q Hemostasis Analyzer for prothrombin time (PT), activated partial aPTT + AT.
thromboplastin time (aPTT), fibrinogen, thrombin time, factors V and VIII, antithrombin
and protein C activity, and von Willebrand factor antigen concentration (vWF:Ag). Total
precision was evaluated by performing 2 runs on 10 separate days, each run consisted of
2 replicates of the same two frozen plasma samples (normal and abnormal) i.e. 40 deter-
Carryover
minations. The results are given as the coefficient of variation (in %). For details, see The sample carryover was investigated by measuring aPTT alterna-
Table 2. tively in 14 aliquots of a pool of normal plasma samples, coded N1 to
Normal sample (n = 40) Pathological sample (n = 40)
N14, and in 4 aliquots of the same pool of normal plasma samples that
was spiked with UFH up to 1.5 IU/mL, final concentration, coded H1 to
PT (sec) 4.4% 4.9%
H4. These samples were processed in a serial way as follows: N1, N2,
aPTT (sec) 3.2% 5.4%
Fibrinogen (g/L) 6.0% 4.8% N3, N4, N5, H1, N6, H2, N7, H3, N8, H4, N9, N10, N11, N12, N13, N14.
Thrombin time (sec) 2.7% 6.0%
Factor V (%) 10.1% (3.0% in sec) 11.2% (3.2% in sec)
Factor VIII (%) 16.0% (3.2% in sec) 16.3% (3.3% in sec) Comparison Study
Antithrombin (%) 7.1% 6.3% Finally, the test results obtained on the Q analyzer using reagents
Protein C (%) 2.7% 5.7%
from the analyzer manufacturer were compared to those obtained
vWF:Ag (%) 7.6% 11.1%
simultaneously on the same plasma aliquot using the current methods
from our laboratory on the ACL TOP analyzer (Instrumentation Labora-
tory, Bedford, MA, USA) using reagents described in the Table 1.
obtain analyte levels in the range from 0% to approximately 100%, de-
Comparisons were performed according to the CLSI EP-09-A2-IR [13].
pending on the test i.e. PT, FV, FVIII, AT, PC and vWF:Ag. Each dilution
was analyzed in duplicate in 5 different days, according to the CLSI
EP06-A guidelines [11]. Reagents and Assay Procedures

Reference Ranges and Control Values
Normal reference ranges were determined for PT and aPTT accord-
ing to the current guidelines [12] by investigating frozen plasma sam-
ples from 25 apparently healthy individuals. The mean normal
prothrombin time (MNPT) and the aPTT control value were calculated
as the arithmetic mean value of the observed clotting times. Similarly,
the normal value and reference range for TT were defined as the mean
value +/− 2 standard deviations (SD) in the plasma from 25 healthy in-
dividuals with normal routine coagulation tests (PT, aPTT, fibrinogen,
and D-dimer).
Throughput
Throughput was evaluated by loading the Q analyzer with 30 sam-
ples obtained from the routine workload of the laboratory, with a mix
of 60% normal and 40% abnormal test results. The effective throughput
was calculated by subtracting the time to obtain the first completed
test panel to that of the last completed panel. Four different test panels
The reagents and the calibration plasma sample (DG-Ref) used on
the Q Hemostasis Analyzer were from the instrument manufacturer
(Diagnostic Grifols).
PT was initiated by adding 100 μL of a rabbit-brain thromboplastin
solution (DG-PT) to 50 μL 37 °C-prewarmed plasma sample without
any incubation. International Normalized Ratio (INR) was calculated
using 1.01 as the lot specific international sensitivity index (ISI) for
the Q analyzer, as indicated by the reagent manufacturer. The MNPT
was determined as previously described (see Reference ranges part).
APTT was initiated by adding 50 μL of a 0.025 mol/L calcium chloride
solution to a mixture of 50 μL aPTT reagent (DG-APTT Synth) and 50 μL
plasma which was previously incubated at 37 °C for 300 sec. The DG-
APTT Synth reagent was a mixture of synthetic phospholipids and
ellagic acid.
Fibrinogen was measured according to Clauss [14] by adding 50 μL of
a thrombin solution (DG-FIB L Human) to 100 μL of 1:7 diluted plasma
in DG-Owren buffer solution.

Table 4
Effective throughput of the Q Hemostasis Analyzer for different test panels including various combination of coagulation i.e. prothrombin time (PT), activated partial thromboplastin time
(aPTT), and fibrinogen (Fg) and chromogenic assays (antithrombin activity: AT). Throughput, expressed as the number of samples evaluated per hour and as the total number of tests
performed per hour, was evaluated by loading the Q Hemostasis Analyzer with 30 fresh plasma samples and determining the time required to obtain all test results.

Test panel Time required to obtain all results (min, sec) Number of samples evaluated per hour Number of tests performed per hour

PT 11’ 43” 154 154

PT + aPTT 27’ 27” 66 132
PT + aPTT + Fg 43’ 49” 42 126
PT + aPTT + AT 48’ 22” 38 114

Table 5
Results of hemostasis tests performed on the Q Hemostasis Analyzer and the current laboratory analyzers i.e. ACL TOP. Analytical comparison of results obtained on the two analyzers was
performed using the Student T-test for paired samples in the case of normally-distributed data (expressed as the mean values with SD) or using the Wilcoxon’s matched pairs T-test in the
case of non-normally distributed data (expressed as the median values with ranges). The correlation between tests results was evaluated using the correlation coefficient in the case of
normally distributed data and using the Spearman’s coefficient of rank correlation in the case of non-normally distributed data.

Test n Q Hemostasis Analyzer Current laboratory method Comparison (p) Correlation (r)

PT (ratio) 107 2.13 (0.88 – 6.66) 2.04 (0.85 – 7.00) NS (0.59) 0.982
INR 58 2.93 ± 1.06 3.10 ± 1.26 b0.0001 0.971
aPTT (sec) 95 37.2 (23.6 – 119.1) 44.5 (25.9 –108.0) b0.0001 0.915
(ratio) 95 1.35 (0.86 – 4.32) 1.35 (0.78 – 6.30) NS (0.68) 0.914
Fibrinogen (g/l) 47 2.79 ± 1.38 2.69 ± 1.33 NS (0.20) 0.933
TT (sec) 164 21.4 (16.7 – 55.3) 24.6 (17.5 – 65.0) b0.0001 0.874
(ratio) 164 1.10 (0.85 2.82) 1.14 (0.81 4.42) b0.0001 0.873
Antithrombin (%) 49 92.2 (12.9 121.9) 99.0 (22.0 132.0) b0.0001 0.925
Protein C (%) 119 70.6 (7.6 – 164.9) 65.4 (3.0 – 170.4) b0.0001 0.989
vWF:Ag (%) 35 137.1 (4.9 – 231.4) 132.4 (6.3 – 251) NS (0.72) 0.862
930 P. Toulon et al. / Thrombosis Research 133 (2014) 927–935

A 0,8

0,6 +1.96 SD
0,54
0,4

PT (ratio) Q - ACL TOP

0,2

0,0 Mean
-0,06
-0,2

-0,4

-0,6 -1.96 SD
-0,66
-0,8

-1,0
0 1 2 3 4 5 6 7 8
Mean of PT (ratio) Q and ACL TOP

B 0,6
+1.96 SD
0,4 0,45

0,2
INR Q - ACL TOP

0,0

Mean
-0,2
-0,17

-0,4

-0,6

-1.96 SD
-0,8
-0,79

-1,0
1 2 3 4 5 6 7 8
Mean of INR Q and ACL TOP

Fig. 1. Comparison performed according to Bland-Altman of the tests results obtained on the Q and the ACL TOP analyzers i.e. prothrombin time (PT in ratio, panel A), international
normalized ratio (INR, panel B), activated partial thromboplastin time (aPTT in ratio, panel C), fibrinogen (in g/L, panel D), thrombin time (in ratio, panel E), antithrombin activity
(in %, panel F), protein C activity (in %, panel G) and von Willebrand factor antigen concentration (vWF:Ag in %, panel H). The dash line indicates the regression line of differences
versus averages.

Thrombin time (TT) was measured by adding 70 μL of a human
thrombin solution (2.0 IU/mL) to 80 μL of plasma which was previously
incubated at 37 °C for 120 sec.
For coagulation factor V (FV) measurement, 50 μL of FV-depleted
plasma (DG-FV) was added to 50 μL of 1:10 diluted plasma in DG-
Owren buffer solution. After a 120 sec-incubation at 37 °C, final mea-
surement was performed after addition of 100 μL of the DG-PT reagent.
For coagulation factor VIII (FVIII) measurement, 50 μL of FVIII-
depleted plasma (DG-FVIII) was added to 50 μL of 1:5 diluted plasma
in DG-Owren buffer solution and 50 μL of the DG-APTT reagent. After
a 180 sec-incubation at 37 °C, final measurement was performed after
addition of 50 μL of a 0.025 mol/L calcium chloride solution.
Antithrombin (AT) activity was assayed using the DG-Chrom AT
chromogenic substrate-based assay, by adding 100 μL of FXa solution
(approx. 2 nKat/mL) to 11 μL of 1:40 diluted plasma in 9.0 g/L NaCl so-
lution. After a 120 sec-incubation at 37 °C, the change in optical density
(OD) at 405 nm was recorded after adding 40 μL of an FXa-specific chro-
mogenic substrate solution.
Protein C (PC) activity was evaluated using the DG-Chrom PC chro-
mogenic substrate-based assay. Briefly, 20 μL of undiluted plasma
were incubated with 70 μL of DG-PC Act, a PC activator solution contain-
ing an Agkistrodon contortrix contortrix snake venom extract, at 37 °C
for up to 180 sec. The change in OD at 405 nm was recorded during
180 sec after adding 70 μL of DG-APC Sust, an activated PC-specific
chromogenic substrate.
The vWF antigen (vWF:Ag) concentration was measured using the
DG-Latex VWF:Ag reagent. Briefly, 130 μL of the latex reagent, a
monoreagent composed of a suspension of latex particles adsorbed
 P. Toulon et al. / Thrombosis Research 133 (2014) 927–935 931

C
1,0
+1.96 SD
0,76
0,5

Mean
0,0

aPTT (ratio) Q - ACL TOP

-0,03

-0,5
-1.96 SD
-1,0 -0,83

-1,5

-2,0

-2,5

-3,0
0 1 2 3 4 5 6
Mean of aPTT (ratio) Q and ACL TOP

D 1,0

0,8
+1.96 SD
0,6 0,68
Fibrinogen (g/L) Q - ACL TOP

0,4

0,2 Mean
0,15
0,0

-0,2

-1.96 SD
-0,4
-0,39

-0,6

-0,8
0 1 2 3 4 5 6
Mean of Fibrinogen (g/L) Q and ACL TOP

Fig. 1 (continued).

with polyclonal antibody against vWF, was added to 45 μL of undiluted
plasma was incubated and the change in OD at 405 nm was recorded
during 300 sec.
Clotting times for FV, FVIII, and fibrinogen were plotted on specific
reference curves which were generated by the analyzer, allowing the
calculation of corresponding analyte concentration. AT and PC activities,
determined using a chromogenic-substrate based assay, were calculated
using a calibration curve generated by the instrument by plotting change
in OD at 405 nm (y axis) against known activities (x axis). Similarly,
vWF:Ag concentrations, evaluated using an immuno-turbidimetric
assay, were determined using the same type of calibration curves as
described above.
For the comparison study, plasma samples were analyzed simulta-
neously on the Q Hemostasis Analyzer using reagent from the instrument
manufacturer and on the ACL TOP analyzer (Instrumentation Laboratory,
Bedford, MA, USA) using IL reagents that were currently used the labora-
tory (Table 1).
Statistical Analysis
Results were expressed as the mean value with SD when the data
were normally distributed. Otherwise, results were expressed as the me-
dian value with range. The Student t-test and the correlation coefficient
(in the case of normally distributed data), the Wilcoxon’s matched pairs
T-test and the Spearman’s coefficient of rank correlation (both in the
case of non-normally distributed data) were used when required. In addi-
tion, the bias analysis, evaluated according to Bland and Altman [15], was
used to test the clinical agreement between the results obtained on the
two analyzers. Statistical analysis was performed using the MedCalc soft-
ware version 12.7.7.0 (MedCalc Software bvba, Mariakerke, Belgium).
932 P. Toulon et al. / Thrombosis Research 133 (2014) 927–935

E 1,0

0,5 +1.96 SD

Thrombin Time (ratio) Q - ACL TOP

0,46

0,0 Mean
-0,10

-0,5 -1.96 SD
-0,65
-1,0

-1,5

-2,0

-2,5
0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0
Mean of Thrombin Time (ratio) Q and ACL TOP

F 15

10 +1.96 SD
8,8
5
Antithrombin (%) Q - ACL TOP

-5 Mean
-6,6
-10

-15

-20 -1.96 SD
-22,0
-25

-30

-35
0 20 40 60 80 100 120 140
Mean of Antithrombin (%) Q and ACL TOP

Fig. 1 (continued).

Results
Precision
Preliminary (within-run) precision was evaluated, for each test, by
evaluating frozen pools of normal and pathological plasmas 20 times
in the same run. As shown in Table 2, the coefficients of variation
(CV%), calculated as the standard deviation divided by the mean value,
was below 4.5% for all tested parameters both in the normal and abnor-
mal ranges except for FVIII in the sample with high FVIII concentration
(mean value = 175.5%) and in the pathological sample (mean value
= 47.0%), with CV% of 6.9% and 11.1%, respectively.
Total precision was evaluated by performing 2 runs on 10 separate
days, each run consisted of 2 replicates of the same two frozen pools
of plasma (normal and pathological) i.e. 40 determinations. As shown
in Table 3, the CV% was below 7% for most of the parameters both in
the normal and pathological ranges except for AT in the normal range
of concentrations (7.1%) and vWF:Ag in both normal and abnormal
ranges of concentrations (7.6% and 11.1%, respectively). The CV% for
FV and FVIII were between 3.0 and 3.3% for the measured clotting
time, but due to the weak slope of the calibration curves, the CV%
were higher when considering test results in percentage activity i.e.
around 10% for FV and 16% for FVIII.
Linearity
The linearity of the calibration curves was determined by using serial
dilutions of a pool of plasma samples in order to obtain parameter levels
in the range from 0% to approximately 100%, depending on the test i.e.
PT, FV, FVIII, AT, PC and vWF:Ag. Each dilution was analyzed in duplicate
 P. Toulon et al. / Thrombosis Research 133 (2014) 927–935 933

G 25

20

+1.96 SD

Protein C (%) Q - ACL TOP

15 15,6

10

Mean
5
5,3

-1.96 SD
-5
-5,1

-10
0 50 100 150 200
Mean of Protein C (%) Q and ACL TOP

H 80

+1.96 SD
60 63,6

40
vWF:Ag (%) Q - ACL TOP

20

Mean
0
2,0

-20

-40

-1.96 SD
-60
-59,7

-80
0 50 100 150 200 250 300
Mean of vWF:Ag (%) Q and ACL TOP

Fig. 1 (continued).

in 5 different days. The linearity of all these tests, evaluated using the
polynomial method, was acceptable for all assays tested with either
no significant non-linearity or a degree of non-linearity less than the
amount of allowable bias for the method.
Reference Ranges and Control Values
Normal reference ranges were determined for PT and aPTT by inves-
tigating plasma samples from 25 apparently healthy individuals. The
MNPT and the aPTT control value were calculated as the arithmetic
mean value of the observed clotting times i.e. 13.2 sec and 27.6 sec re-
spectively. The reference ranges, calculated as the mean value +/− 2
SD in that population, were set between 11.4 and 15.0 sec for PT and be-
tween 22.8 and 32.4 sec for aPTT, with corresponding patient-to-control
ratios in the range from 0.86 and 1.14 for PT and from 0.83 to 1.17 for
aPTT. Similarly, the TT control value (19.6 sec) was defined as the arith-
metic mean value of the observed clotting times obtained in the plasma
from 25 healthy individuals with normal routine coagulation tests. The
reference range was set between 16.9 and 22.3 sec with corresponding
patient-to-control ratios in the range from 0.86 and 1.14.
Throughput
Throughput was evaluated for different test panels by loading the Q
Hemostasis Analyzer with 30 fresh plasma samples. When the single PT
test mode was chosen, the time to obtain the first test result was 3.5
min. and the measured throughput was 154 PT per hour. Different test
panels including two to three different tests were also evaluated and
the Q throughput varied in the range from 114 to 154 tests per hour de-
pending of the tested panel (Table 4).
934 P. Toulon et al. / Thrombosis Research 133 (2014) 927–935
Carryover
The sample carryover was investigated by measuring aPTT alterna-
tively in 14 aliquots of a pool of normal plasma samples and in 4 aliquots
of the same pool of normal plasma samples that was spiked with
unfractionated heparin up to 1.5 IU/mL final concentration (for details,
see Materials and Methods section). As the result, no prolongation of
normal plasma aPTT was detected in alternating measurements of
these two plasma samples (data not shown).
Correlation with the ACL TOP Analyzer
By evaluating plasma samples obtained from patients (n variable de-
pending of the test), the results of the different clotting, chromogenic
and immunologic assays obtained on the Q analyzer were highly corre-
lated with those obtained on the ACL TOP with r in the range from 0.862
to 0.989 (Table 5). Most of the test results obtained on the two analyzers
was significantly different, except for PT (ratio), fibrinogen and vWF:Ag
levels. However, these analytical differences had no clinical relevance in
most of the cases when data were compared according to Bland-Altman
(Fig. 1). Actually, only few samples would have been classified as defi-
cient for any specific analyte on the Q Hemostasis Analyzer and normal
on the ACL TOP or conversely, and these discrepant results
corresponded to levels near the lower limit of the normal range (not
shown).
Evaluation of Plasma Samples with Abnormal Optical Characteristics
Eighteen hemolyzed plasma samples, with plasma hemoglobin con-
centrations up to 175 mg/L (normal value below 50 mg/L) and 15 icter-
ic plasma samples, with total bilirubin concentrations up to 51.3 μmol/L
(normal range: 3–17 μmol/L) were investigated. PT and aPTT test re-
sults obtained on the Q were similar to those obtained on the ACL
TOP, another optical clot detection-based analyzer with reading at 671
nm. The same applied to the three lipemic samples available during
the analyzer evaluation period (not shown).
Practical Experience
No problem was encountered during the 5 month-evaluation of the
Q instrument which was found to be easy-to-use by the technicians. The
software was user-friendly with clear presentation of data on a large
scaled touch-screen, despite the small size of lettering.
Discussion
The Q Hemostasis Analyzer is a multiparameter analyzer that was de-
signed to simultaneously perform clotting, colorimetric and immuno-
logic assays. Its analytical performances were found to be similar to
those of other fully automated coagulation analyzers such as the STAR
[16], the CA7000 [17,18], the MDA-180 [19] or the ACL TOP [20,21],
with intra- and inter-assay precision (evaluated as the CV) below 6%
for most of the studied parameters. Our results were congruent with
those previously reported in an external evaluation of the Q analyzer
[22]. In patients’ plasmas, the test results obtained on the Q were highly
correlated with those obtained on the ACL TOP. We demonstrated that
the minor but significant analytical differences between test results ob-
tained using the two analyzers had no clinical relevance (Bland-Altman
analysis). This latter point is of critical importance since analytical or
statistical significance is purely numerical whereas clinical significance
is a judgment based on a variety of clinical decision-making criteria
that includes, but are not limited to, whether results are likely to affect
management of individual patients. Moreover, only few samples
would have been misclassified as deficient for any specific analyte
with only one of the two analyzers. The vast majority of such discrepant
results corresponded to values near the lower limit of the normal range
of the various parameters and therefore had no clinical relevance. The
explanation for such discrepancies is likely to be multifactorial and re-
lated not only to differences in the assay sensitivity, in the calibrator
used in the assay, in the analyzer characteristics such as the curve fitting
abilities or its optical performance but also to the within-laboratory test
performance, as already observed in external quality assessment
schemes for different parameters [23]. As the Q is a photo-optical clot
detection-based analyzer, we specifically investigated plasma samples
with optical abnormalities. The potential impact of moderate hemolysis
and icterus on the coagulation tests results (PT and aPTT) was highly
unlikely, since test results obtained on the Q were similar to those ob-
tained on the ACL TOP, another optical clot-detection-based analyzer,
with OD monitoring at 671 nm. Such a wavelength was chosen on the
ACL TOP to avoid any interference with hemoglobin and bilirubin [20].
Even if congruent results were obtained on both analyzers in the
three lipemic samples available during the analyzer evaluation period,
further investigations are needed to conclude that high triglyceride con-
centration, as encountered in patients on intravenous lipid administra-
tion, have no potential impact on the coagulation test results performed
on the analyzer. The Q Hemostasis Analyzer throughput fulfilled the re-
quirements of medium-sized laboratories workload for both routine
and specialized testing as well as for emergency situations. For larger
laboratory requiring higher throughput, the connectivity capability of
the Q analyzer could enable to set up a network of two or more instru-
ments managed by a single interface [24]. Finally, the Q analyzer was
found to have potential advantages which included, but are not limited
to, a high throughput, continuous loading capabilities for samples, re-
agents and disposables, cap-piercing capability and a lack of any
sample-to-sample carryover. In addition, adding assays from other
manufacturers to the Q analyzer is very easy, thanks to a user-friendly
dedicated screen, in which the test name, unit(s), sample and reagents
volumes, sample dilution ratio if applicable, incubation and acquisition
times could be easily defined.
In conclusion, the Q Hemostasis Analyzer was found to be precise and
reliable not only for routine coagulation testing but also for the determi-
nation of single factor activities, chromogenic substrate-based and im-
munologic assays. Together with the short duration of the daily
maintenance which was limited to 15 min, its throughput, its continu-
ous loading capabilities for samples, reagents and disposables, and its
user-friendly software makes it a convenient choice for medium scale
hemostasis laboratories.
Conflict of Interest Statement
The authors have no conflicts of interests.
Acknowledgements
We are indebt to the technicians from the Department of Hematolo-
gy (Pasteur University Hospital, Nice, France) and we want to thank par-
ticularly Sylvain Buvat, Christine Sigaud and Monique Lemaire for their
expert technical assistance during the evaluation of the Q Hemostasis
Analyzer.
References
[1] Walenga JM, Fareed J. Automation and quality control in the coagulation laboratory.
Clin Lab Med 1994;14:709–28.
[2] Aiach M, Leon M, Michaud A, Capron L. Adaptation of synthetic peptide substrate-
based assays on a discrete analyzer. Semin Thromb Hemost 1983;9:206–16.
[3] Amiral J, Adalbert B, Adam M. Application of enzyme immunoassays to coagulation
testing. Clin Chem 1994;30:1512–6.
[4] Gardiner C, Kitchen S, Dauer RJ, Kottke-Marchant K, Adcock D. Recommendations for
evaluation of coagulation analyzers. Lab Hematol 2007;12:32–8.
[5] Gardiner C, Adcock DM, Carrington LR, Kottke-Marchant K, Marlar RA, McGlasson
DL, et al. CLSI Document H57-A. Protocol for the evaluation, validation and imple-
mentation of coagulometers: approved guidelines, vol. 28, n°4. Wayne, PA, USA:
Clinical and Laboratory Standards Institute; 2008.
 P. Toulon et al. / Thrombosis Research 133 (2014) 927–935
[6] WHO. Guidelines for thromboplastin and plasma used to control oral anticoagulant
therapy. WHO Technical Report series, 889; 1999. p. 64–93.
[7] Walker ID. Blood collection and sample preparation: pre-analytical variation. In:
Jespersen J, Bertina RM, Haverkate F, editors. Laboratory techniques in thrombosis -
A manual. ECAT assay procedures, 2nd revised ed. Dordrecht, The Netherlands:
Kluwer Academic Publishers; 1999. p. 21–8.
[8] Polack B, Schved J-F, Boneu B, on behalf of the ‘Groupe d'Etude sur l'Hémostase et la
Thrombose’ [GEHT]. Preanalytical recommendations of the ‘Groupe d'Etude sur
l'Hémostase et la Thrombose’ (GEHT) for venous blood testing in hemostasis labora-
tories. Haemostasis 2001;31:61–8.
[9] Krouwer JS, Cembrowski GS, Tholen DW. CLSI Document EP10-A3. Preliminary eval-
uation of quantitative clinical laboratory measurement procedures; approved guide-
lines, 3rd ed., vol. 26, n°34. Wayne, PA, USA: Clinical and Laboratory Standards
Institute; 2006.
[10] Tholen DW, Kallner A, Kennedy JW, Krouwer JS, Meier K. NCCLS Document EP5-A2.
Evaluation of precision performance of quantitative measurement methods; ap-
proved guideline, 2nd ed., vol. 24, n°25. Wayne, PA, USA: Clinical and Laboratory
Standards Institute; 2004.
[11] Tholen DW, Kroll M, Astles JR, Caffo AL, Happe TM, Krouwer JS, et al. NCCLS Docu-
ment EP6-A. Evaluation of the linearity of quantitative measurement procedures: a
statistical approach; approved guidelines, vol. 23, n°16. Wayne, PA, USA: Clinical
and Laboratory Standards Institute; 2003.
[12] Marlar RA, Cook J, Johnston M, Kitchen S, Machin SJ, Shafer D, et al. CLSI Document
H47-A2. One-Stage Prothrombin time (PT) test and activated partial thromboplastin
time (APTT) test: approved guideline, 2nd ed., vol. 28, n°20. Wayne, PA, USA: Clinical
and Laboratory Standards Institute; 2008.
[13] Krouwer JS, Tholen DW, Garber CC, Goldschmidt HMJ, Kroll M, Linnet K, et al. CLSI
Document EP09-A2-IR. (interim revision) , Method comparison and bias estimation
935
using patient samples; approved guidelines, 2nd ed., vol. 30, n°17. Wayne, PA, USA:
Clinical and Laboratory Standards Institute; 2010.
[14] Clauss A. Gerinnungsphysiologische Schnellmethode zur Bestimmung des Fibrino-
gens. Acta Haematol 1957;17:237–46.
[15] Bland JM, Altman DJ. Statistical methods for assessing agreement between two
methods of clinical measurement. Lancet 1986;1:307–10.
[16] Flanders MM, Crist R, Safapour S, Rodgers GM. Evaluation and performance charac-
teristics of the STA-R coagulation analyzer. Clin Chem 2002;48:1622–4.
[17] Dom-Beineke A, Dempfle CE, Bertsch T, Wisser H. Evaluation of the automated
coagulation analyzer Sysmex CA-7000. Thromb Res 2005;116:171–9.
[18] Fischer F, Appert-Flory A, Jambou D, Toulon P. Evaluation of the automated coagula-
tion analyzer Sysmex® CA 7000. Thromb Res 2006;117:721–9.
[19] Mohammed A, Mehrabani PA, Coombs R, Molesworth L, Favaloro EJ. The MDA-180
coagulation analyser: a laboratory evaluation. Pathology 1997;29:176–83.
[20] Appert-Flory A, Fischer F, Jambou D, Toulon P. Evaluation and performance character-
istics of the automated coagulation analyzer ACL TOP. Thromb Res 2007;120:733–43.
[21] Milos M, Herak D, Kuric L, Horvat I, Zadro R. Evaluation and performance character-
istics of the coagulation system: ACL TOP analyzer-HemosIL reagents. Int J Lab
Hematol 2009;31:26–35.
[22] Kitchen S, Woolley A. Evaluation of the Q analyzer (Diagnostic Grifols), a new
cap-piercing fully automated coagulometer with clotting, chromogenic and
immunoturbidimetric capability. Blood Coagul Fibrinolysis 2013;24:28–34.
[23] Preston FE, Kitchen S, Srivastava A. External quality assessment in hemostasis: its im-
portance and significance. In: Kitchen S, Olson JD, Preston FE, editors. Quality in labo-
ratory hemostasis and thrombosis. Chichester, UK: Wiley-Blackwell; 2009. p. 51–62.
[24] Pages J, Cases R, Torres S. Scaling up Q hemostasis analyzer to high throughput
laboratories. J Thromb Haemost 2011;9(Suppl. 2) [Abstract P-MO-223].