Lupus anticoagulant: performance of the tests as recommended by the latest ISTH guidelines

Journal of Thrombosis and Haemostasis, 9: 1776 1783 DOI: 10.1111/j.1538-7836.2011.04420.x ORIGINAL ARTICLE Lupus anticoagulant: performance of the tests as recommended by the latest ISTH guidelines J. SWADZBA,* T. IWANIEC,* M. PULKA,* B. DE LAAT, à P.G.DEGROOT and J. M U S I A L * *Department of Medicine, Jagiellonian University Medical College, Krakow, Poland; Department of Clinical Chemistry and Haematology, University Medical Center, Utrecht; àsanquin Research, Amsterdam; and Synapse, CARIM, Maastricht University Medical Center, Maastricht, the Netherlands To cite this article: Swadzba J, Iwaniec T, Pulka M, de Laat B, de Groot PG, Musial J. Lupus anticoagulant: performance of the tests as recommended by the latest ISTH guidelines. J Thromb Haemost 2011; 9: 1776 83. Summary. Objectives: Lupus anticoagulant (LA) is clinically the most relevant among all antiphospholipid antibody tests. Recently, new guidelines for LA detection were published. The objective of this retrospective cohort study was to compare tests recommended under these guidelines with other methods used for LA detection. Methods: The study group consisted of 336 subjects suffering from various autoimmune diseases. We used activated partial thromboplastin time (aptt), diluted Russell viper venom time (drvvt) and diluted prothrombin time (dpt) tests for LA detection together with a ratio between sensitive and insensitive aptt reagent. We also tested if LA was dependent on b 2 glycoprotein I (b 2 GPI) using one of the recently described methods. Results: All LA tests performed were associated with a history of thrombosis. The highest odds ratio (OR) for thrombosis was found for b 2 GPI-dependent LA but sensitivity was low (OR = 8.4; specificity/sensitivity = 98%/ 15%). All LA tests showed a much stronger association with thrombosis than with pregnancy failure. Conclusions: LA tested by aptt and/or drvvt (at least one out of two tests positive), as recommended by the guidelines, was associated less strongly with a history of thrombosis (OR = 4.1) than either of these tests separately (OR = 5.0 and 4.3, respectively). With both tests positive (Ôdouble LA positivityõ) the association with thrombosis was stronger (OR = 6.5) compared with only one positive test. In fact, Ôdouble LA positivityõ, detected by combinations of any of the tests studied, was markedly associated with a history of thrombosis. Keywords: guidelines, lupus anticoagulant, double LA positivity, b2gpi-dependent LA. Correspondence: Jakub Swadzba, Department of Medicine, Jagiellonian University Medical College, Skawinska 8, 31-006 Kraków, Poland. Tel.: +48 12 430 5314; fax: +48 12 430 5314. E-mail: swadzba@diag.pl Received 10 March 2011, accepted 9 June 2011 Introduction The term Ôlupus anticoagulantõ (LA) was used for the first time in 1972 by Feinstein and Rapaport [1] to describe a non-specific coagulation inhibitor found 20 years earlier in some patients with systemic lupus erythematosus (SLE) [2]. In 1963 it was found that in spite of circulating anticoagulant some of those patients suffered from thrombosis [3]. In the 1980s, it became clear that LA belongs to a group of autoantibodies directed against negatively charged phospholipids [4]. Then, in the 1990s, antiphospholipid antibody protein cofactors, b 2 glycoprotein I (b 2 GPI) and prothrombin, were described [5 7]. Although sensitive and quantitative ELISAbased methods were developed to detect antiphospholipid antibodies, LA detected by coagulometric tests has been shown to be more associated with thrombosis [8], and for this reason remained as an indispensable, clinically relevant laboratory method in evaluating antiphospholipid syndrome (APS). APS is defined as the co-occurrence of antiphospholipid antibodies (apl) with characteristic clinical symptoms [9]. From the beginning, LA and anticardiolipin antibodies (acl) were included among laboratory criteria of the syndrome [9,10]. In 2004, antibodies against b 2 GPI (ab 2 GPI) formed the last addition to the criteria [11]. A set of guidelines for the detection of LA were published in 1991 and 1995 [12,13] and then revised in 2009 [14]. The revision has been introduced for several important reasons. LA detection is troublesome (a three step procedure), poorly standardized and its laboratory accuracy and clinical utility varies, igniting a lot of controversies [15]. Pre-analytical (e.g. quality of sample collection, centrifugation, and temperature of storage) and analytical (e.g. large number of methods and reagents used) variables strongly influence the final results. As a consequence, several national and international inter-laboratory surveys showed unacceptable differences in LA test results between various participating laboratories [16,17]. The rates of false-positive and false-negative results remained relatively high. Recently, two tests have been recommended by recent guidelines for LA detection based on practicality and global

Lupus anticoagulant: performance of the tests 1777 experience: diluted Russell viper venom time (drvvt) and activated partial thromboplastin time (aptt) [14,18 21]. In the past, other tests were used for this purpose, for example kaolin clotting time (KCT) [22], silica clotting time (SCT) [23], diluted prothrombin time (dpt) [16], aptt sensitive/insensitive ratio (aptt ratio) [24] and Textarin/Ecarin clotting time [25]. Since then, their use has been discouraged mainly because of the limited experience rather than their poor performance. Interestingly, according to one systematic literature review, the risk of thrombosis appeared to be independent of the laboratory tests used for LA identification [8]. It has been shown that ab 2 GPI are better predictors of thrombotic complications than antibodies directed against prothrombin [26]. For this reason, various attempts have been made to specifically detect b 2 GPI-dependent LAs. Two methods of b 2 GPI-dependent LA identification were described [27,28]. The first studies indicate that these b 2 GPI-dependent LAs show a superior association with thrombotic complications than LAs caused by other antibodies [29,30]. From a clinical point of view, any LA results should always be considered in the context of a full laboratory apl profile comprising of acl and ab 2 GPI, as isolated LA positivity may be frequently found in subjects without any clinical symptoms [14,31]. Starting from the revised guidelines for LA detection, we decided to compare retrospectively performance of the recommended LA tests with other frequently used tests to detect LA in a large population of patients with generalized autoimmune diseases. Subjects A group of 336 subjects suffering from generalized autoimmune diseases was studied: 283 women and 53 men, aged 21 77 (mean: 43 years). Systemic lupus erythematosus (SLE) was diagnosed in 235 patients based on the presence of at least four American College of Rheumatology (ACR) criteria [32], and SLE-like disease in 44 (three ACR criteria, out of which one was immunologic). In 32 patients, APS was not accompanied by any features of another autoimmune disease: primary antiphospholipid syndrome (PAPS). Out of the SLE and SLElike group, 80 patients were diagnosed with a secondary antiphospholipid syndrome (SAPS). All the patients have been under continuous care of the Outpatient Clinic for Autoimmune Diseases, Department of Medicine, Jagiellonian University Medical College, Krakow, Poland. A detailed history (a uniform set of questions) was taken from all the patients by a physician and medical records were carefully analyzed. Objective data confirming typical clinical complications and laboratory test abnormalities of APS were required. All thrombotic episodes had to be confirmed by the imaging techniques [ultrasonography (USG), computed tomography (CT) and/or nuclear magnetic resonance (NMR)]. Pregnancy morbidity considered typical for APS according to recent updated classification criteria [11] included: (i) one or more unexplained deaths of a morphologically normal fetus 10 weeks of gestation, or (ii) one or more premature births < 34 weeks of gestation, or (iii) three or more unexplained consecutive spontaneous abortions < 10 weeks of gestation. Pregnancy complications were confirmed by an obstetrician based on available medical records. Based on history, medical records and objective imaging techniques, clinical signs typical for APS were identified in 180 patients (53.6%). Thrombotic complications occurred in 142 cases (deep venous thrombosis: 84, arterial thrombosis: 49, thrombosis in both vascular beds: 9). A history of obstetrical complications was reported by 73 out of 214 women (34%), who had ever been pregnant. Co-existence of thrombosis and obstetrical complications was identified in 35 women. All clinical episodes have happened in the past more than 3 months, but < 5 years before blood samples were collected. Eighty-four patients were treated with oral anticoagulants (OA) at the time of blood collection. All patients with an international normalized ratio (INR) higher than three were excluded from the study. Thirty-eight patients had an INR ratio between 2 and 3, with the remainder having an INR lower than 2 on the day of blood collection. One hundred healthy subjects, from the city of Krakow, constituted a population sample. It then was matched by gender and age formed a control group. Controls served to establish upper limits (99 percentile) of the normal values for methods used in the study. Methods Blood was collected into plastic tubes containing 0.109 M sodium citrate as an anticoagulant, centrifuged (2000 g, 15 min at room temperature) and the resulting plasma was transferred to another plastic tube for a second spin. After the second spin, plasma was aliquoted and immediately frozen ()70 C). Frozen plasma was thawed in a water bath at 37 C immediately before testing. LA was detected according to the new guidelines using two recommended tests and a three-step procedure [14]. PTT LA (Diagnostica Stago, Asnieres sur Seine, France) and the DVV test (American Diagnostica, Stamford, CT, USA) were used for screening and mixing steps, and Staclot LA (Diagnostica Stago) and DVV Confirm (American Diagnostica) reagents for the confirmation. Moreover, three other tests were performed: dpt, the sensitive/insensitive aptt ratio (aptt ratio), together with b 2 GPI dependence of the LA in all samples positive for aptt-based LA. The cut-off values were set at the 99 percentile established for plasmas from 100 healthy donors. The methods used are described below: 1 LA detection using an aptt test (aptt-dependent LA): I. PTT LA test sensitive for LA (Diagnostiga Stago), normal values: up to 45.6 s. II. Only samples with positive results in step I: PTT LA mixing studies (1:1 with normal pooled plasma). Negative result if the value is in the normal range (up to 45.6 s). III. Only samples with positive results in step II: Staclot LA (Diagnostiga Stago), normal values: up to 8 s (difference

1778 J. Swadzba et al between aptt and aptt with addition of hexagonal phospholipids). 2 LA detection by a drvvt test (drvvt dependent LA): I. DVV test (American Diagnostica), normal values: up to 46.1 s. II. Only samples with positive results in step I: DVV test mixing studies (1:1 with normal pooled plasma). Negative result if the value is in the normal range (up to 46.1 s). III. Only samples with positive results in step II: DVV confirm (American Diagnostica), normal values: DVV test/dvv confirm ratio lower than 1.3. 3 aptt sensitive/insensitive ratio (aptt ratio; LA aptt ratio dependent): I. PTT LA test sensitive for LA (Diagnostiga Stago), Normalvalues:upto45.6s. II. aptt test insensitive for LA (Pathromtin SL; Siemens, Munich, Germany), Normal values: up to 40 s. III. Calculation of ratio (aptt ratio): aptt sensitive/insensitive, normal values up to 1.4. 4 LA detection by a dpt test (dpt-dependent LA): I. Innovin (Siemens) diluted 1:200 in OwrenÕs veronal buffer and incubated 7 min, Normal values up to 54.0 s. II. Only samples with positive results in step I: dpt mixing studies 1:1 with normal pooled plasma. Normal values up to 54.0 s. 5 Detection of b 2 GPI-dependent LA [30]: I. Only aptt LA positive samples were studied (method 1 positive). II. PTT LA mixing studies (1:1 with normal pooled plasma) were repeated with the addition of 50 ll of cardiolipin. III. Ratio of PTT LA mixing studies (1:1 with normal pooled plasma) and pooled plasma were calculated (1). IV. Ratio of PTT LA mixing studies (1:1 with normal pooled plasma) with addition of 50 ll of cardiolipin and pooled plasma with addition of 50 ll of cardiolipin were calculated (2). V. Calculation of the difference of ratios (1 and 2), Results > 0.05 show b 2 GPI dependent LA. The combination of LA aptt-dependent and b 2 GPIdependent LA tests were excluded from the calculation because of the bias (the b 2 GPI-dependent LA test is a part of the apttdependent LA group). LA was referred as Ôdouble LA positivityõ when it was detected by at least two methods. The tests were run on a Behring Coagulation Timer (BCT) and Fibrintimer (Siemens). Odds ratios (ORs) with 95% confidence intervals (CIs) for previously diagnosed clinical symptoms of APS in relation to the positivity of LA were calculated using the v 2 -test. Pregnancy failure ORs were calculated only in a group of women who were ever pregnant (n = 214; 73 with and 141 without obstetrical symptoms of APS). Results According to the current guidelines for using aptt and drvvt tests, LA was identified in 95 patients. APTTdependent LA was detected in 85 patients (aptt-dependent LA group). The drvvt-dependent LA group included 64 patients. Both tests were positive in 54 cases. The latter group was defined as showing double LA positivity. LAbasedondPTandLAbasedonaPTTratiotestswere found in a slightly higher number of subjects (106 and 101 subjects, respectively) than LA based on the recently recommended tests. When b 2 GPI dependence of the aptt positive LA samples (85 cases) was analyzed, it was found to be positive in 25 patients (26%), which represents only 7.5% of the population studied. The association of all five tests with the history of thrombosis (arterial and/or venous) and pregnancy complications is shown in Table 1. In terms of clinical utility, positive results of all tests were associated with thrombosis (ORs range: 3.5 8.4). The strongest association and high specificity (98%) was found for b 2 GPIdependent LA, but its sensitivity was low (15%). Maybe because of its low sensitivity, b 2 GPI-dependent LA showed no association with pregnancy complications (Table 1). Table 1 An association of single lupus anticoagulant (LA) tests with clinical symptoms of antiphospholipid syndrome (APS) Method of LA detection Thrombosis (n = 142) Specificity/sensitivity Venous thrombosis (n = 93) Specificity/sensitivity Arterial thrombosis (n = 57) Specificity/sensitivity Pregnancy failure (n = 73) Specificity/sensitivity aptt 5.0 (2.9 8.6) 5.0 (2.9 8.5) 2.2 (1.2 4.1) 2.2 (1.1 4.3) 87/43 84/48 78/39 84/30 drvvt 4.3 ( 7.8) 3.5 (2.0 6.3) 2.1 (1.1 4.1) 2.3 (1.1 5.2) 90/32 87/34 83/30 90/21 dpt 6.0 (3.6 10.0) 4.4 (2.6 7.3) 2.6 (1.4 4.6) 84/53 78/55 73/49 aptt ratio 3.5 (2.2 5.8) 3.7 (2.2 6.1) 2.1 (1.2 3.8) 2.1 (1.1 4.0) 81/45 78/51 73/44 79/36 b 2 GPI-dependent LA 8.4 (2.8 24.9) 4.4 (1.9 10.2) 98/15 96/16 OR, odds ratio; CI, confidence interval; aptt, activated partial thromboplastin time; drvvt, diluted Russell viper venom time; dpt, diluted prothrombin time; b 2 GPI, b 2 glycoprotein I.

Lupus anticoagulant: performance of the tests 1779 Table 2 An association of a single lupus anticoagulant (LA) test with thrombosis in a group of patients with no vitamin K antagonist (VKA) treatment Method of LA detection Patients without VKA treatment (n = 252) aptt 2.9 (1.5 5.8) dvv 2.9 (1.4 6.1) dpt 2.5 (1.3 4.7) aptt ratio 2.2 (1.2 4.0) b 2 GPI-dependent LA 5.8 (1.7 20.0) OR, odds ratio; CI, confidence interval; aptt, activated partial thromboplastin time; drvvt, diluted Russell viper venom time; dpt, diluted prothrombin time; b 2 GPI, b 2 glycoprotein I. Diluted PT-dependent LA showed a stronger association with the history of thrombosis than aptt- or drvvtdependent LAs (OR = 6.0 vs. 5.0 and 4.3, respectively). On the other hand, LA based on the ratio between sensitive and insensitive aptt reagents showed a weaker association with APS clinical symptoms than other LA tests. As oral anticoagulants could affect differently various LA tests, we performed additional calculations in a subgroup of patients not treated with oral anticoagulants at the time of blood collection (Table 2) Obviously, the results showed weaker associations because of the lower number of patients, and the lower number of confirmed clinical events (all of the 84 OA-treated patients suffered from thrombosis in the past). b 2 GPI-dependent LA remained the most strongly associated with thrombosis, but dpt-dependent LA showed a weaker association with thrombosis than the LA tests recommended by the current guidelines. Additional sub-analysis was performed depending on the patientsõ age (under 45 vs. 45 years or more). The first group consisted of 217 subjects, with 81 patients suffering from thrombosis, whereas the second consisted of 119 patients (61 with thrombosis). The OR for thrombosis and LA tests did not differ significantly between these groups with the exception of b 2 GPI-dependent LA (data not shown). Interestingly, all of the 11 patients older than 45 years and positive for b 2 GPIdependent LA suffered from thrombosis. When venous and arterial events were analyzed separately, aptt-dependent LA showed the strongest association with venous thrombosis (OR = 5) whereas dpt-dependent LA showed the strongest association with arterial thrombosis (OR = 2.6). DRVVT-dependent LA correlated stronger than other tests with pregnancy complications (OR = 2.3), but, in general, the associations of any LA with these complications were weaker than those with thrombosis (Table 1). The definition of LA based on the recommendations requires positivity by either aptt and/or drvvt tests. Interestingly, the risk of thrombosis based on the apttpositive result only, or solely on the drvvt test was in both situations higher than the risk based on positivity of either of the tests combined (OR; 5.0; 4.3; and 4.1, respectively) (Table 1, Fig. 1). The advantage of the latter (aptt and/or drvvt) lays only in their higher sensitivity (45% vs. 43% for aptt-dependent LA and 32% for drvvt-dependent LA). Moreover, LAs detected by recently recommended tests were certainly not better than LAs detected by the dpt and/or aptt ratio (OR = 4.1 vs. 5.1 respectively) (Fig. 1). In fact, thrombosis was most strongly associated with LA detected by the dpt test and/or the drvvt test (OR = 6.3) (Fig. 1). When venous and arterial events were analyzed separately, the association of venous thrombosis was the strongest with LA detected by aptt and/or dpt tests and that of arterial thrombosis for LA detected by drvvt and/or dpt tests (Fig. 1). There needs to be more research performed to identify how we should define the presence of a strong LA as opposed to the weak ones. Taking advantage of the multiple LA tests performed in the same patient, we also analyzed patients Pair of tests; and /or positivity Thrombosis n = 141 Venous thrombosis n = 93 Arterial thrombosis n = 57 Pregnancy complications n = 73 1.9 1.6 1.8 1.7 1.9 1.6 2.0 aptt - drvvt 4.1 4.0 2.2 aptt - dpt 5.8 4.9 2.6 aptt - aptt ratio 3.5 3.7 2.0 drvvt - dpt 6.3 4.6 2.9 drvvt - aptt ratio 3.4 3.2 dpt - aptt ratio aptt ratio - β 2 GPIdep. LA 5.1 3.7 4.7 3.7 2.3 drvvt - β 2 GPIdep. LA 4.8 4.2 2.3 2.1 dpt - β 2 GPIdep. LA 6.1 4.7 2.5 1.7 0 1 2 3 4 5 6 7 8 9101112 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 0 1 2 3 4 5OR (Cl) Fig. 1. An association between antiphospholipid syndrome (APS) clinical symptoms and lupus anticoagulant (LA) positivity (positivity in either one of the two tests).

1780 J. Swadzba et al Double LA positivity Thrombosis n = 141 Venous thrombosis n = 93 Arterial thrombosis n = 57 Pregnancy complications n = 73 aptt - drvvt 6.5 5.0 2.2 3.1 aptt - dpt aptt - aptt ratio drvvt - dpt drvvt - aptt ratio dpt - aptt ratio 5.6 5.5 4.2 5.3 4.6 4.7 5.2 3.5 4.7 3.8 2.3 1.9 1.9 2.8 2.8 2.5 3.1 2.8 aptt ratio - β 2 GPIdep. LA 13.2 6.3 3.1 drvvt - β 2 GPIdep. LA 10.6 3.6 2.3 4.0 dpt - β 2 GPIdep. LA 10.0 4.2 3.1 0 5 10 20 30 40 50 60 012 4 6 8 10 12 14 16 18 0 1 2 3 4 5 6 7 8 012 4 6 8 1012141618202224 OR (Cl) Fig. 2. An association between antiphospholipid syndrome (APS) clinical symptoms and lupus anticoagulant (LA) positivity in both tests (double LA positivity). according to the number of positive tests. Those with two positive LA tests (also referred to as Ôdouble LA positivityõ)are presented in Fig. 2. Generally, when two tests were positive (double LA positivity), the association with clinical symptoms of APS was much stronger (Fig. 2) than when only one test was positive. When both recently recommended tests were positive, the risk for thrombosis was quite high (OR = 6.5) but still close to a single positivity with dpt (OR 6.0; Table 1). Overall, Ôdouble LA positivityõ was characterized by a lower sensitivity but better specificity than LA detected according to the recent guidelines. ÔDouble LA positivityõ was also analyzed for all other possible combinations of the tests used (Fig. 2). The strongest association between thrombosis and any two positive tests were seen for the aptt ratio and b 2 GPI-dependent LA (OR 13.2; specificity 99%). Any combination which included b 2 GPIdependent LA showed a superior association with thrombosis, far exceeding other possible combinations. Finally, we analyzed associations of APS clinical symptoms with LA Ôtriple positivityõ. The results showed improved specificity at the further expense of sensitivity, but ORs remained similar (results not shown). The highest specificity for thrombosis that is equal to 100% was shown for the combination of the following three: b 2 GPI-dependent LA, the aptt ratio and drvvt. Discussion The recently published Subcommittee for Standardisation (SSC) guidelines for LA detection [14] provided a big step forward in attempting to improve quality and reliability of LA testing. The guidelines recommend aptt and drvvt tests as the two preferred coagulometric tests for a screening step, based on, by far, the largest research performed with these tests. Other specific requirements include silica as an activator and low phospholipid content. In the present study, we compared four different screening tests for LA detection in 336 subjects with generalized autoimmune diseases. Moreover, a fifth test was performed with the aptt-positive LA patients aimed at the detection of b 2 GPI dependence of the LA phenomenon. Our study has shown that positive results of all performed tests were associated with the risk of thrombosis in the past and/or pregnancy complications (Table 1). However, the strength of these associations has differed depending on the test used and/ or their combinations. Several of our findings concur with the new guidelines or suggest a need for further studies to improve LA detection. First, we showed that the performance of dpt and the aptt ratio showed a comparable association with clinical symptoms (Table 1) to the tests recommended by the recent guidelines. In fact, dpt LA positivity showed a higher association with thrombosis (OR = 6.0) with a good sensitivity/specificity ratio in all the patients studied. Good performance of the dpt test has already been reported by others [16,33]. The test was criticized mainly because of the variability of various thromboplastin reagents used and the limited number of manufacturers [14]. This might be true, but knowing that some reagents perform quite satisfactorily and may improve the sensitivity/ specificity of LA detection, one may suggest performing a comparative study. The performance of the dpt test become less impressive when OA-treated patients were excluded from the calculations. It indicates that a dpt test is more sensitive to the influence of OA than other tests, resulting in many falsepositive results of the mixing (a second step) studies. It should be mentioned at this point that the dpt method does not have a third confirmatory step. For this reason, the use of dpt for LA detection should be probably limited to only patients not treatedwithoa. The other method used in the present study for LA detection was based on the use of sensitive and insensitive aptt reagents, as described previously [24]. This method is not recommended

Lupus anticoagulant: performance of the tests 1781 because it does not follow a classic three-step procedure. We agree that it should not probably be recommended for a routine LA detection. It could be useful, however, in specialized laboratories, when results are not clear, to make a decision; for example, when the prolongation of aptt could be owing to reasons other than LA (e.g. samples from patients treated with heparin). Second, aptt-positive LA should be followed by a test detecting b 2 GPI dependency. It increases substantially the likelihood of thrombosis in such patients (OR = 8.4), as has been already reported [29,30]. We would suggest it, however, as a second line test because of its low sensitivity. Prospective studies are necessary to determine a real risk associated with of b 2 GPI-dependent LA for future clinical APS complications. It would be also important to perform a test detecting b 2 GPIdependence of LA regardless of the method of initial LA detection. Tests were similarly associated with thrombosis, both in older (more than 45 years) and in younger patients. Multivariate analyzes in properly powered studies would be required to assess an additional independent influence of classic venous and arterial thrombosis risk factors on thrombotic events in LA-positive subjects. For many years, at least two screening tests were suggested to detect LA [13]. Any single test is not sensitive for all different types of LA. The different tests are based on slightly different principles and phospholipids. Recently, drvvt and aptt combined have been recommended to be performed to detect LA, as they were the most widely described in the medical literature. These tests were also routinely used in the past decade in our laboratory [34,35] showing acceptable results. It has been shown that a further increase in the number of the screening tests above two increased the risk of false-positive results to an unacceptable level [14]. Interestingly, our results showed that the chance of experiencing thrombosis in the past (OR) was higher for any of these single tests than for combining both tests (Table 1 and Fig. 1). It may suggest that false-negative findings resulting from the test added prevail over putative increased specificity and predictive values. Overall, two screening tests combined did not provide better results than that of a single test. If we accept a notion that two tests could increase the detection of different clinically significant antiphospholipid antibodies with anticoagulant properties, a combination of two tests could be recommended as a result of our main research study (Fig. 1). Interestingly, because of the cost and labor requirements, several recent studies were still using only one screening test [36]. However, this is another important reason to perform more than one screening test. This allows for the identification of a group of patients who have two tests positive (Ôdouble LA positivityõ). The ORs for thrombosis are higher in the groups of patients with double LA positivity (Fig. 2) than in groups of patients with only one positive test (Table 1) and the groups of patients with one or two tests positive (Fig. 1). Here, positivity for b 2 GPI-dependent LA combined with positive LA detected by other methods show especially a strong association with thrombotic complications. Based on our study, Ôtriple LA positivityõ did not increase the strength of these associations. LA has a stronger association with recurrent fetal losses than other apl [37], but not as strong as with the thrombosis. In the present study, the association was even weaker because, in recent years, most of the pregnant women were treated with aspirin or/ and low-molecular-weight heparin (LMWH) and many of them had successful pregnancies. In APS, the pathogenesis of pregnancy failure may be different than that of thrombosis and this may be one of the reasons why b 2 GPI-dependent LA was not significantly associated with pregnancy loss in our study whereas it was so highly associated with thrombosis. APL is quite heterogeneous. We are still looking for precise immunologic and/or functional characteristics of apl, which could determine their clinical relevance. A high-risk profile of apl is widely discussed [38 40]. One such property is certainly the LA effect which combined with medium or high titers of acl and ab 2 GPI antibodies (only IgG subtype), under the name of Ôtriple positivityõ, might form an especially high-risk profile for thrombosis and pregnancy complications. In such a triplet, LA is known as the most important antibody connected with higher risk of thrombosis than the others [33,41]. However, LA as a single test is not always a risk factor for thrombosis [31]. Positive b 2 GPI-dependent LA may form just another element of such a high-risk profile. The test showed very high specificity of almost 100%, not seen in any other LA tests, but at the expense of a very low sensitivity. Interestingly, another recently performed test was able to detect specific apl, namely, antibodies against domain I of b 2 GPI that shows also poor sensitivity [42]. The proper place of these tests in the diagnostic workup of APS requires further prospective trials. Based on our results and those of the otherõs [14,29,30,42], a high-risk apl-positive population, most probably, includes patients with a multiple positivity for apl (acl, ab 2 GPI and LA), those with Ôdouble LA positivityõ, and patients with positive for antibodies against domain I or b 2 GPI-dependent LA. It is obvious that based on one main research study performed in a single population using one set of reagents, it is not possible to prove that a given test is better than the other. Also the present study was performed with a group of patients suffering mostly from generalized autoimmune diseases, and the results should not be directly extrapolated to the general population or patients with primary antiphospholipid syndrome. However, the results show that the clinical results of the tests chosen in the recent guidelines are at least equal to some other tests that have been excluded. Therefore, we recommend that at least LA detection by dpt with certain thromboplastins and tests showing b 2 GPI dependency of the LA effect require further research studies as potentially useful laboratory tests in APS. If clinically relevant, they may find their place in new guidelines for LA detection in the future. Conclusions 1 All LA tests positively correlated with the previous presence of thrombosis.

1782 J. Swadzba et al 2 All LA tests were associated much stronger with thrombosis than with pregnancy failure. 3 b 2 GPI-dependent LA as a single test was the most strongly associated with thrombosis but showed very low sensitivity. 4 Double LA positivity detected by all tests performed was strongly associated with the history of thrombosis. Acknowledgement This study was supported in part by a research grant No K/ ZDS/000610 from the Jagiellonian University Medical College. Disclosure of Conflict of Interest The authors state that they have no conflict of interest. References 1 Feinstein DI, Rapaport SI. Acquired inhibitors of blood coagulation. Prog Haemost Thromb 1972; 1: 75 95. 2 Conley CL, Hartmann RC. A haemorrhage disorder caused by circulating anticoagulant in patients with disseminated lupus erythematosus. J Lab Clin Invest 1952; 31: 621 2. 3 Bowie EJW, Thompson JH, Pascuzzi CA, Owen CA. Thrombosis in systemic lupus erythematosus despite circulating anticoagulants. JLab Clin Med 1963; 62: 416 30. 4 HarrisEN,GharaviAE,BoeyML,PatelBM,LoizouS,HughesGR. Anticardiolipin antibodies: detection by radioimmunoassay and association with thrombosis in lupus erythematosus. Lancet 1983; 2: 1211 4. 5 GalliM,ComfuriusP,MaassenC,HemkerHC,DeBaetsMH,van Breda-Vriesman PJC, Barbui T, Zwaal RFA, Bevers EM. Anticardiolipin antibodies (ACA) directed not to cardiolipin but to plasma protein cofactor. Lancet 1990; 335: 1544 7. 6 McNeil HP, Simpson RJ, Chesterman CN, Krilis SA. Anti-phospholipid antibodies are directed against a complex antigen that includes the lipid binding inhibitor of coagulation: b2-glycoprotein I (apolipoprotein H). Proc Natl Acad Sci USA 1990; 87: 4120 4. 7 Bevers EM, Galli M, Barbui T, Comfurius P, Zwall RF. Lupus anticoagulant IgGÕs (LA) are not directed to phospholipids only, but to a complex of lipid bound human prothrombin. Thromb Haemost 1991; 66: 629 32. 8 Galli M, Luciani D, Bertollini G, Barbui T. Lupus anticoagulants are stronger risk factors for thrombosis than anticardiolipin antibodies in the antiphospholipid syndrome: a systemic review of the literature. Blood 2003; 101: 1827 32. 9 Harris EN. Syndrome of the black swan. Br J Rheumatol 1987; 26: 324 6. 10 Wilson WA, Gharavi AE, Koike T, Lockshin MD, Branch DW, Piette J-C, Brey R, Derksen R, Harris EN, Hughes GRV, Triplett DA, Khamashta MA. International consensus statement on preliminary classification criteria for definite antiphospholipid syndrome. Arthritis Rheum 1999; 42: 1309 11. 11 Miyakis S, Lockshin MD, Atsumi T, Branch DW, Brey RL, Cervera R, Derksen RHWM, de Groot PG, Koike T, Meroni PL, Reber G, Shoenfeld Y, Tincani A, Vlachoyiannopoulos PG, Krilis SA. International consensus statement on an update of the classification criteria for definite antiphospholipid syndrome (APS). JThrombHaemost 2006; 4: 295 306. 12 Exner T, Triplett DA, Taberner D, Machin SJ. Guidelines for testing and revised criteria for lupus anticoagulants. Thromb Haemost 1991; 65: 320 2. 13 Brandy JT, Triplett DA, Alving B, Scharer I. Criteria for diagnosis of lupus anticoagulants: an update. On behalf of the Subcommittee on Lupus Anticoagulant/Antiphospholipid Antibody of the Scientific and Standardisation Committee of the ISTH. Thromb Haemost 1995; 74: 1185 90. 14 Pengo V, Tripodi A, Reber G, Rand JH, Ortel TL, Galli M, de Groot PG. Update of the guidelines for lupus anticoagulant detection. J Thromb Haemost 2009; 7: 1737 40. 15 Tripodi A. Testing for lupus anticoagulants: all that a clinician should know. Lupus 2009; 4: 291 8. 16 Arnout J, Meijer P, Vermylen J. Lupus anticoagulant testing in Europe: an analysis of results from the first European Concerted Action on Thrombophilia (ECAT) survey using plasmas spiked with monoclonal antibodies against human beta2-glycoprotein I. Thromb Haemost 1999; 81: 929 34. 17 Pengo V, Biasiolo A, Gresele P, Marongiu F, Erba N, Veschi F, Ghirarduzzi A, de Candia E, Montaruli B, Testa S, Barcellona D, Tripodi A. Survey on lupus anticoagulant diagnosis by central evaluation of positive plasma samples. Thromb Haemost 2007; 5: 925 30. 18 PengoV,BiasioloA,RampazzoP,BroccoT.dRVVTismoresensitive than KCT or TTI for detecting lupus anticoagulant activity of anti-beta2-glycoprotein I autoantibodies. Thromb Haemost 1999; 81: 256 8. 19 Urbanus RT, Derksen RH, de Groot PG. Current insight into diagnostics and pathophysiology of antiphospholipid syndrome. Blood Rev 2008; 2: 93 105. 20 Tripodi A, Biasiolo A, Chantarangkul V, Pengo V. Lupus anticoagulant(la)testing:performance of clinical laboratories assessed by national survey using lyophilized affinity purified immunoglobulin with LA activity. Clin Chem 2003; 49: 1608 14. 21 Lambert M, Ferrard-Sasson G, Dubucquoi S, Hachulla E, Prin L, Hatron PY, Goudemand J, Boffa MC, Caron C. Diluted Russell vipervenom time improves identification of antiphospholipid syndrome in a lupus anticoagulant-positive patient population. Thromb Haemost 2009; 101: 577 81. 22 GalliM,FinazziG,BeversEM,BarbuiT.Kaolinclotingtimeand dilute Russell viper venom timedistinguish between prothrombin dependent and beta -2 glycoprotein-i dependent antiphospholipid antibodies. Blood 1995; 86: 617 23. 23 DragoniF,MinottiC,PalumboG,FaillaceF,RediR,BongarzoniV, Avvisati G. As compared to kaolin clotting time, silica clotting time is a specific and sensitive automated method for detecting lupus anticoagulant. Thromb Res 2001; 101: 45 51. 24 Ames PR, Iannaccone L, de Lasio R, Brancacio V. Improved confirmation of weak lupus anticoagulants by employing sensitive and insensitive reagents to the lupus anticoagulant. Blood Coagul Fibrynolisis 2001; 12: 563 7. 25 TriplettDA,StockerKF,UngerGA,BarnaLK.TheTextarin/Ecarin ratio: a confirmatory test for lupus anticoagulants. Thromb Haemost 1993; 70: 925 31. 26 Swadz ba J, de Clerck LS, Stevens WJ, Bridts CH, van Cotthem KA, Musial J, Jankowski M, Szczeklik A. Anticardiolipin antibodies, antib2-glycoprotein I, antiprothrombin antibodies and lupus anticoagulant in patients with systemic lupus erythematosus with a history of thrombosis. J Rheumatol 1997; 24: 1710 5. 27 Simmelink MJA, Derksen RHWM, Arnout J, de Groot PG. A simple method to discriminate between beta2-glycoprotein I and prothrombin dependent lupus anticoagulants. J Thromb Haemost 2003; 1: 740 7. 28 Pengo J, Biasiolo A, Pegoraro C, Iliceto S. A two-step coagulation test to indentify anti beta2-glycoprotein I lupus anticoagulants. Thromb Haemost 2004; 2: 702 7. 29 De Laat B, Derksen RHWM, Urbanus RT, Roest M, de Groot PG. Beta2 glycoprotein-i dependent lupus anticoagulant highly correlates with thrombosis in the antiphospholipid syndrome. Blood 2004; 104: 3598.

Lupus anticoagulant: performance of the tests 1783 30 De Laat B, Derksen RHWM, Reber G, Musial J, Swadzba J, Bozic B, Cucnic S, Regnault V, Forastiero R, Woodhams BJ, de Groot PG. An international multicentre-laboratory evaluation of a new assay to detect specifically lupus anticoagulants dependent on the presence of antibeta2-glycoprotein autoantibodies. J Thromb Haemost 2011; 9: 149 53. 31 Pengo J, Biasiolo A, Gresele P, Marongiu F, Erba N, Veschi F, Ghirarduzzi A, Barcellona D, Tripodi A. A comparison of lupus anticoagulant positive patients with clinical picture of antiphospholipid syndrome and those without. Arterioscler Thromb Vasc Biol 2007; 27: 309 10. 32 Hochberg MC. Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1997; 40: 1725. 33 Horbach DA, van Oort E, Donders CJM, Derksen RHWM, de Groot PG. Lupus anticoagulant is a strongest risk factor for both venous and arterial thrombosis in patient with systemic lupus erythematosus. Comparison between different assays for the detection of antiphospholipid antibodies. Thromb Haemost 1996; 76: 916 24. 34 Musial J, Swadzba J, Motyl A, Iwaniec T. Clinical significance of antiphospholipid protein antibodies. Receiver operating characteristics plot analysis. J Rheumatol 2003; 30: 723 30. 35 Swadzba J, Iwaniec T, Szczeklik A, Musial J. Revised classification criteria for antiphospholipid syndrome and the thrombotic risk in patients with autoimmune diseases. JThrombHaemost2007; 5: 1883 9. 36 Petri M. Update on anti-phospholipid antibodies in SLE: the HopkinsÕ Lupus Cohort. Lupus 2010; 19: 419 23. 37 Opatrny L, David M, Kahn SR. Association between antiphospholipid antibodies and recurrent fetal loss in women without autoimmune diseases: a metaanalysis. J Rheumatol 2006; 33: 2214 21. 38 Galli M, Reber G, de Moerloose P, de Groot PG. Invitation to a debate on the serological criteria that define the antiphospholipid syndrome. J Thromb Haemost 2008; 6: 399 401. 39 Swadzba J, Musial J. More on: the debate on antiphospholipid syndrome classification criteria. J Thromb Haemost 2009; 7: 501 2. 40 PengoV,BanzatoA,BisonE,DenasG,PadayattilJoseS,RuffattiA. Antiphospholipid syndrome: critical analysis of the diagnostic path. Lupus 2010; 19: 428 31. 41 Cervera R, Piette JC, Font J, Khamashta MA, Shoenfeld Y, Camps MT, Jacobsen S, Lacos G, Tincani A, Galeazzi M, Meroni PL, Derksen RH, de Groot PG, Baleva M, Mosca M, Bombardieri S, Housiau F, Gris JC, Quere I, Hachulla E, et al. Antiphospholipid syndrome: clinical and immunologic manifestations and patterns of disease expression in a cohort of 1000 patients. Arthritis Rheum 2002; 46: 1019 27. 42 De Laat B, Pengo V, Pabinger I, Musial J, Voskuyl AE, Bultink IE, Ruffatti A, Rozman B, Kveder T, de Moerloose P, Boehlen F, Rand J, Ulcova-Gallova Z, Mertens K, de Groot PG. The association between circulating antibodies against domain I of beta2-glycoprotein I and thrombosis: an international multicenter study. J Thromb Haemost 2009; 7: 1767 73.