Thrombosis Research (2004) 114, 547--552

intl.elsevierhealth.com/journals/thre

Detection of procoagulant phospholipid interfering

in tests for lupus anticoagulant
Thomas Exner *, Joyce Low

C/-Haematology Department, St. Vincents Hospital, Sydney, Victoria Street, Darlinghurst,

NSW 2010, Australia

Received 25 May 2004; received in revised form 29 May 2004; accepted 1 June 2004
Available online 14 July 2004

KEYWORDS ABSTRACT
Lupus anticoagulant/
procoagulant Excess platelets shorten most clotting tests for lupus anticoagulant (LA). Often it is
phospholipid; not clear if a shortened, normal or slightly prolonged result in a test masks a weak LA
Activated platelets in combination with activated platelets, which express procoagulant phospholipid
(PPL). Our aim was to investigate a new LA-insensitive factor Xa-activated clotting
time (XACT) test for detecting PPL in plasma specimens submitted for LA testing. In
most clotting tests for PPL, specimens are mixed with human platelet-free plasma
(PFP) to correct for factor defects. Such tests are usually very sensitive to
prolongation by LA, which act against PPL-human clotting factor complexes. We
found that phospholipid-free plasma from pigs could be used instead of human
platelet-free plasma as substrate plasma without reducing sensitivity of XACT to
PPL. However, the pig plasma-based system was significantly less affected by most
LA. Activated platelets were detectable despite the presence of most LA. Since some
LA still had significant prolonging effect on the XACT despite the use of pig plasma
we investigated this further. ELISAs for IgG and IgM anti-h2GP1 and anti-prothrombin
antibodies were carried out on 23 specimens. We did not find that LA in plasmas
displaying anti-prothrombin antibodies had less prolonging effect on the test based
on pig plasma than that using human platelet-free plasma. Similarly, there were no
subtyping trends apparent among results from anti-h2GPI-positive samples. Our
results do not support the concept that anti-prothrombin-dependent LA might be
more species specific than anti-h2GPI-dependent LA.
A 2004 Elsevier Ltd. All rights reserved.

Introduction

The interfering role of platelets in lupus anticoag-

* Corresponding author. Tel.: +61-2-9498-4296; fax: +61-2-
ulant (LA) testing systems is somewhat complex.
9845-7038. Fresh normal platelets do not seem to interfere at
E-mail address: exner@optushome.com.au (T. Exner). all in some tests. The PRP recalcification time

0049-3848/$ - see front matter A 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.thromres.2004.06.001
548 T. Exner, J. Low

described by Gastineau et al. [1] used fresh plate-
let-rich plasma samples and was considered by
many as a highly sensitive test for LA. On the other
hand, washed frozen platelets have been used in
many labs to bypass [2] and neutralise LA in tests
such as the platelet neutralisation procedure de-
veloped by Triplett et al. [3].
This issue is resolved by the observation that
platelets only neutralise or bypass LA when they
have been disrupted by freeze thawing or have been
otherwise activated so that their procoagulant
phospholipid (PPL) has become exposed externally.
In unstimulated platelets membrane asymmetry is
maintained by an aminophospholipid translocase
system [4] so that procoagulant phosphatidylserine
is not normally available on the surface for clotting
interactions.
Clotting tests for LA are usually based on phos-
pholipid-deficient systems, which are often sensi-
tive to shortening by activated platelets. The
current recommendations for LA testing already
make provisions for reducing platelet counts in test
samples to below 10  109 l 1 [5].
The degree of interference varies across differ-
ent types of test; kaolin clotting time (KCT) usually
being most shortened by PPL, followed by dilute
Russell’s viper venom (DRVVT), tissue thromboplas-
tin inhibition (TTI) time and activated partial
thromboplastin time (APTT) tests. This depends
also on the concentration and type of PPL used in
the test system (T. Exner, unpublished). Often it is
not clear if a shortened, normal or slightly pro-
longed result in one of the more PPL-sensitive tests
may mask a weak LA in combination with activated
platelets, which express PPL. Conversely, LA inter-
fere with most clotting test systems for PPL, which
are also depleted in phospholipid content.
Our aim was to investigate a new test, which
might be more specific for detecting activated
platelets or microparticles in plasma specimens
submitted for LA testing [6].
Materials and methods
Generally, blood samples were collected by veni-
puncture into Vacutainer (Becton-Dickinson, USA)
3.2% citrate tubes. Platelet-rich plasmas were har-
vested after centrifuging at 400  g for 5 min.
Platelet-poor plasmas were obtained after 20 min
at 2500  g. Platelet-free plasmas (PFP) were

Table 1 Details of some test plasma (where known)

Clinical note APTT TTI mix RVV mix RVV mix Screen/confirm
(s) ratio screen ratio confirm ratio ratio
1. PNP normal 32 1.0 1.0 1.0 1.0
2. LA1 F35, SLE > 1.25
3. LA2 M42, FII inhibitor >1.25
4. LA3 M55, SLE,OAT >1.25
5. LA4 SLE >1.25
6. LA5 M54, ITP >1.25
7. LA6 F25, SLE >1.25
8. LA7 F >1.25
9. LA8 M, OAT 143 >4.2 3.44 1.49 2.31
10. LA9 F 51 3.42 2.06 1.18 1.75
11. L10 F >1.25
12. L11 M 3.35 2.32 1.13 2.05
13. L12 M >1.25
14. L13 M 1.18 1.59 1.02 1.56
15. L14 M 33 1.24 1.01 1.23
16. L15 M, OAT 3.77 2.02 1.03 1.96
17. L16 M >1.25
18. L17 F 34 1.01 1.25 0.99 1.26
19. L18 F 57 2.3 2.3 1.25 1.84
20. L19 M >1.25
21. L20 M 51 1.25
22. L21 F 40 3.41 2.50 1.20 2.08
23. L22 F 62 >1.25
24. L23 F 46 3.58 1.91 1.14 1.68
Pooled normal plasma designated as PNP.
Patients on warfarin therapy are designated ‘‘OAT’’. All patients were known to have LA from prior testing
protocols. SLE stands for systemic lupus erythematosus. ITP for idiopathic thrombocytopenic purpura.
Detection of procoagulant phospholipid interfering in tests for lupus anticoagulant 549

and IgM antibodies were kindly provided by Diag-

nostica Stago and the tests were carried out exactly
as described in the kits inserts. The mean + 2 stan-
dard deviation cutoff levels were assumed to be 10
GAU/ml and 10 MAU/ml for anti-h2GPI phenotypes
and known to be 10 GAU/ml and 10 MAU/ml for the
anti-prothrombin phenotypes.

Results

Specimens

Table 1 summarises the information available on

plasmas used in this project. All except LA6 had
Figure 1 XACT calibration curve. Effect of platelet
been double centrifuged before freezing to mini-
concentration added to platelet-free human plasma on
XACT tests based on either platelet-free human plasma mise platelet contamination. LA6 plasma had been
(.) or phospholipid depleted porcine plasma (E). obtained after plasmapheresis to prolong a difficult
pregnancy and had been frozen with some platelets
present. Lupus inhibitor potencies as shown by the
obtained by centrifuging again at 16,000  g for 5 final DRVVT ratio varied widely and the specimens
min in a microfuge and filtering plasma through a
0.22-Am cellulose acetate filter (Sartorius, Ger-
Table 2 Table shows factor Xa-activated clotting
many). Platelet counts were done on a Coulter S
time results (XACT) on test samples prepared by
Plus STKS (Beckman/Coulter, USA). mixing patient plasmas 1:1 with either normal human
A panel of 23 plasmas with LA activity according plasma or porcine plasma
to ISTH criteria [5] was used. Most of these speci-
Patient plasma XACT/human XACT/porcine
mens were derived from citrated blood and had
been double centrifuged and frozen at 80 C. 1. PNP 46.0 48.0
DRVVT tests were carried out on 50 Al 1:1 mixes 2. LA1 78.4 58.0
3. LA2 91.4 50.4
of test and normal plasma using 50 Al of either LA
4. LA3 77.2 62.0
screen and LA confirm reagents (Gradipore, 5. LA4 68.8 51.8
Frenchs Forest, Sydney). APTT tests were carried 6. LA5 73.2 70.0
out using 50 Al test plasmas, 50 Al of AutoAPTT 7. LA6 45.8 40.8
reagent (Biomerieux, France), 4 min preincubation 8. LA7 65.0 47.4
and 50 Al 25 mM calcium chloride. Tissue thrombo- 9. LA8 59.6 58.8
plastin inhibition (TTI) tests were carried out using 10. LA9 69.6 49.6
50 Al test plasmas, 50 Al of Innovin (Dade-Behring, 11. L10 74.0 54.2
USA) diluted 1:200 in saline and 50 Al of calcium 12. L11 102 71.6
chloride. These routine LA tests were all carried 13. L12 51.8 47.8
out on a Staclot Instrument (Diagnostica Stago, 14. L13 54.2 49.0
15. L14 60.4 57.0
Asnieres, France).
16. L15 58.2 60.0
XACT test was carried out as previously described 17. L16 72.8 71.8
[6]. Essentially plasma samples of 25 Al were mixed 18. L17 55.8 51.8
and prewarmed for 2 min at 37 C with 25 Al of 19. L18 61.0 55.4
platelet-free human or phospholipid depleted por- 20. L19 65.0 70.0
cine plasma. Then 100 Al of XACT reagent was 21. L20 57.6 57.2
added and the time for clotting determined. Tests 22. L21 124 60.4
were carried out on an ACL300 machine (Instru- 23. L22 60.8 50.6
mentation Laboratories [IL]/Beckman/Coulter, 24. L23 65.0 50.0
Milan, Italy) in research mode 2. The XACT reagent Both human and porcine substrate plasmas used for
contained Polybrene to neutralise up to 5 U/ml of mixing contained trace added freeze-thawed PRP
regular heparin. equivalent to platelet concentration of 3  109 l 1.
Asserachrom ELISA kits for anti-h2GPI-binding Baseline XACT results without the added PRP were 55 s
IgG and IgM antibodies and anti-prothrombin IgG for human and 58 s for porcine plasma.
550
were quite heterogeneous in clinical background,
though typical of what a large haematology depart-
ment would receive.
Calibration
To establish the responsiveness of XACT tests based
on either human or pig phospholipid deficient plas-
mas to platelets, specimens were prepared con-
taining varying dilutions of freeze-thawed platelet-
rich plasma (initial platelet count 250  109 l 1)
from 1/50 to 1/200 in human platelet-free plasma.
XACT tests were carried out on these samples using
both pig phospholipid depleted plasma and human
platelet-free substrate plasmas. Results obtained
are shown in Fig. 1.
It is apparent that XACT tests based on either
human PFP or pig phospholipid-free plasmas
showed similar responsiveness (slope) to human
PRP. Tests in pig plasma tended to give slightly
longer results (by approximately 2--3 s) than when
using human platelet-free substrate plasma, but
sensitivities were similar.
XACT tests on LA samples spiked with
platelets
We next wanted to check the sensitivity of the two
tests to various LA plasmas. A small fixed amount of
Figure 2 XACT results on various LA plasmas spiked with freeze-thawed PRP with initial platelet count of 3  109 l 1.
Results arranged in ascending order. Tests were carried out using either platelet-free human plasma (striped columns)
or phospholipid depleted porcine plasma (solid black columns). Horizontal dashed lines show upper limit of tests
without added platelets: 55 s for human platelet-free plasma, 58 s for phospholipid-depleted pig plasma.
T. Exner, J. Low
freeze-thawed platelets was included in both test
systems to reduce variation in the clotting times,
which may have arisen from traces of procoagulant
phospholipid in the test samples.
Lupus inhibitor plasmas were mixed with an
equal volume of normal platelet-free plasma to
which had been added a small but consistent con-
centration of normal freeze-thawed platelet-rich
plasma (2% of an initial 300  109 l 1). Parallel
studies were carried out with LA specimens added
to an equal volume of phospholipid depleted pig
plasma spiked with the same quantity of freeze-
thawed platelet-rich plasma. (Final level of plate-
let material in both substrate plasmas was thus
3  109 l 1).
All these specimens (50 Al) were then tested
with the XACT reagent (100 Al). Results obtained
are shown in Table 2. XACT times were also ar-
ranged in ascending order to produce Fig. 2.
It is apparent from Table 2 and Fig. 2 that XACT
results with LA present were generally shorter in
the pig plasma system than in the human PFP
system. Thus, LA had less prolonging effect in pig
plasma than in human plasma.
If a horizontal cutoff line is drawn corresponding
to zero added platelets for each system (i.e. 58 s
for the porcine system and 55 s for the human
platelet-free plasma), it is apparent that 15 of
the 23 LA samples are below this line in the test
with pig plasma, whereas only 2 of them are below
Detection of procoagulant phospholipid interfering in tests for lupus anticoagulant 551

the line in the test with human substrate plasma.

Thus, the LA interfere much more with detection of
trace PPL in the human plasma system and much
less when pig plasma is present.

ELISA tests for subtyping the LA samples

The LA specimens were tested for anti-h2GPI and

anti-human prothrombin-binding IgG and IgM anti-
bodies by ELISA. Results obtained are shown in
Table 3.
A wide range of results was obtained confirm-
ing significant heterogeneity among the 23 sam-
ples. Ten specimens displayed elevated levels of
both anti-h2GPI and anti-prothrombin antibodies.
There were three showing only anti-h2GPI and
also three showing only anti-prothrombin antibod-
ies. Six patients did not have any elevated ELISA
result in any category and these were mainly
weaker LA. Figure 3 Lupus inhibitors and significance of anti-
XACT results obtained in the human and porcine h2GPI typing. Correlation between XACT results using 1:1
systems were plotted against each other to obtain mixes of various LA plasmas with either pig or human
phospholipid depleted plasmas. R2 = 0.2556. Size and
the correlation diagrams in Figs. 3 and 4. These
style of points is indicative of strength and isotype of
anti-h2GPI antibodies assessed by ELISA. (Strong IgG>50
AU/ml = ; medium IgG 10--50 AU/ml = ; weak or
Table 3 Table shows results from ELISAs for anti- negative IgG < 10 AU/ml = ; strong IgM>50 AU/ml = ;
bodies against h-2-glycoprotein I (h2GPI) and against medium IgM 10--50 AU/ml = ; weak or negative IgM < 10
human prothrombin AU/ml = ).
IgG/anti- IgM/anti-h IgG/anti- IgM/anti-
h2GPI 2GPI prothrombin prothrombin
1. PNP 0 2.5 3 0 figures also attempt to show the IgG and IgM iso-
2. LA1 200 19 60 2 type of the ELISA result on the source plasma. The
3. LA2 11 7 120 0 larger the point, the stronger the ELISA result. It is
4. LA3 31 6.5 23 0 apparent that the strongest ELISA results either for
5. LA4 0 35 62 18
anti-h2GP1 or anti-prothrombin do not distribute in
6. LA5 0 2 35 0
7. LA6 2 5 3 0
any consistent pattern. Therefore, it is not possible
8. LA7 0 1 5 0 to say that LA in plasmas with stronger anti-pro-
9. LA8 0 4.5 5 75 thrombin antibodies have less effect when tested in
10. LA9 0 7.5 5 115 a pig plasma system than those from plasmas with
11. L10 14 170 14 100 stronger anti-h2GPI titres.
12. L11 46 200 100 105
13. L12 0 1 2.5 0
14. L13 0 3 5 0
15. L14 0 1.5 2 8
16. L15 65 2.5 14 6
Discussion
17. L16 210 48 4 5
18. L17 0 12 2 3 The results obtained in this study showed that
19. L18 0 44 34 10 XACT tests based on human or pig plasmas free
20. L19 215 52 5 4 of platelets or deficient in procoagulant phospho-
21. L20 0 1 1 0 lipid gave similar results on normal human plate-
22. L21 0 165 1 29 let-free plasma and had similar sensitivities to
23. L22 31 1.5 1 0 procoagulant phospholipid expressed by freeze-
24. L23 28 1 120 0 thawed platelets.
Results are expressed in G or M AU/ml as provided by However, most lupus anticoagulant plasmas had
suppliers of the kits (Diagnostica Stago, France). less prolonging effect in the system using pig plas-
Cutoffs were all at 10 AU/ml. ma. Only two LA gave slightly longer XACT in the pig
552
Figure 4 Lupus inhibitors and significance of anti-
prothrombin antibody titre. Correlation between XACT
results using 1:1 mixes of various LA plasmas with either
pig or human phospholipid depleted plasmas. R2 = 0.2556.
Size and style of points is indicative of strength of anti-
prothrombin antibodies as assessed by ELISA. (Strong
IgG>50 AU/ml = ; medium IgG 10--50 AU/ml = ; weak or
negative IgG < 10 AU/ml = ; strong IgM>50 AU/ml = ;
medium IgM 10--50 AU/ml = ; weak or negative IgM < 10
AU/ml = ).
system than in when mixed with human plasma.
This appears to support the hypothesis that some
LA may be species specific. In particular, it has
been suggested that prothrombin-dependent LA
may be less active when animal (bovine) prothrom-
bin is substituted for human prothrombin in test
systems [7].
Analysis of anti-phospholipid subtypes in this
group of patients was unable to confirm this
hypothesis. Unfortunately, plasmas from pigs and
cows contain only about half the level of pro-
thrombin as in human plasma [8] and perhaps this
was just not enough antigen to overcome some of
these LA. Interestingly, pig plasma contains ap-
proximately five times as much factor V as human
plasma [8].
There have been many past attempts to subclas-
sify LA, which have not been successful [9,10]. Two
new simple methods have recently been suggested.
That by Pengo et al. [11] suggests that lower
calcium levels may specifically prolong dilute PT
and dilute RVVT tests on h2 GPI-dependent LA. The
method of Simmelink et al. [12] suggests that
prolonged APTTs on h2GPI-dependent LA may be
specifically shortened by addition of cardiolipin
T. Exner, J. Low
vesicles. Subtyping LA may be inherently problem-
atic anyway because, as found in this group of 23
typical LA patients, most have a mix of anti-pro-
thrombin and h2GPI-binding antibodies. Only six
showed any preference for one antigen or the
other.
Since the pig plasma-based system was less
affected by LA, it should be possible to detect
procoagulant phospholipid even in the presence of
weaker LA. This may be useful in clinical conditions
where platelet activation may sometimes be com-
bined with lupus inhibitors.
Acknowledgements
We are grateful to Dr. Barry Woodhams of
Diagnostica Stago for providing Asserachrom ELISA
kits and to the haematologists at St. Vincents
Hospital, Darlinghurst for access to leftover anon-
ymous LA samples.
References
[1] Gastineau DA, Kazmier FJ, Nichols WL, et al. Lupus anti-
coagulant: an analysis of the clinical and laboratory fea-
tures of 219 cases. Am J Hematol 1985;19:265--73.
[2] Howard MA, Firkin BG. Investigations on the lupus-like by-
passing activity of platelets. Thromb Haemost 1983;50:
775--7.
[3] Triplett DA, Brandt JT, Kaczor D, et al. Laboratory diagnosis
of lupus inhibitors: A comparison of the tissue thromboplas-
tin inhibition procedure with a new platelet neutralisation
procedure. Am J Clin Pathol 1983;79:678--82.
[4] Sims PJ, Wiedmer T. Unravelling the mysteries of phospho-
lipid scrambling. Thromb Haemost 2001;86:266--75.
[5] Brandt JT, Triplett DA, Alving B, Scharrer I. Criteria for the
diagnosis of lupus anticoagulants: an update. Thromb Hae-
most 1995;74:1185--90.
[6] Exner T, Joseph J, Low J, Connor D, Ma D. A new activated
factor Xa-based clotting method with improved specificity
for procoagulant phospholipid. Blood Coagul Fibrinolysis
2003;14:773--9.
[7] Rao LV, Huang AD, Rapaport SI. Differences in the interac-
tions of lupus anticoagulant IgG with human and bovine
prothrombin. Thromb Haemost 1995;73:668--74.
[8] Lewis JH. In Comparative Hemostasis in Vertebrates. New
York: Plenum Press; 1996. p. 340.
[9] Exner T. Similar mechanisms of various lupus anticoagu-
lants. Thromb Haemost 1985;53:15--8.
[10] Exner T. Methods for subtyping lupus anticoagulants. Lupus
1998;7(Suppl. 2):s103--6.
[11] Pengo V, Biasolo A, Pegoraro C, Iliceto S. A two step coag-
ulation test to identify anti-beta 2 glycoprotein I lupus anti-
coagulants. J Thromb Haemost 2004;2:702--7.
[12] Simmelink MJ, Derksen RH, Arnout J, de Groot PG. A simple
method to discriminate between beta 2-glycoprotein I- and
prothrombin-dependent lupus anticoagulants. J Thromb
Haemost 2003;1:740--7.