Anti-prothrombin antibodies combined with lupus anti-coagulant activity is an essential risk factor for venous thromboembolism in patients with systemic lupus erythematosus

Authors


Junzo Nojima, Central Laboratory for Clinical Investigation, Osaka University Hospital, 2–15 Yamadaoka, Suita, Osaka 565-0871, Japan. E-mail: nojima@hp-lab.med.osaka-u.ac.jp

Abstract

Anti-prothrombin antibodies (anti-prothrombin) and anti-β2-glycoprotein I antibodies (anti-β2-GP I) are the most common and characterized anti-phospholipid antibodies (aPL) detected using specific enzyme-linked immunosorbent assay (ELISA) systems. Recently, lupus anti-coagulant (LA) activity detected by a phospholipid-dependent coagulation assay was reported to be associated with anti-prothrombin and/or anti–β2-GP I. Here we show that the co-existence of IgG anti-prothrombin and LA activity might be an essential risk factor for venous thromboembolism (VTE) in patients with systemic lupus erythematosus (SLE). We examined not only the levels of antibodies to prothrombin and anti-β2-GP I (both IgG and IgM isotypes) using an ELISA system, but also LA activity detected using both diluted Russell's viper venom time (dRVVT) and STACLOT LA test in 124 patients with SLE. The SLE patients were divided into four groups according to the results of ELISA and LA assay results for each aPL: group A, ELISA+ and LA+ group B, ELISA+ and LA; group C, ELISA and LA+ group D, ELISA and LA. Regarding IgG anti-prothrombin, the prevalence of VTE was significantly higher in group A (16/35 cases, 45·7%, P < 0·001, Fisher's exact probability test) than in the other groups (B, 2/30, 6·7%; C, 1/22, 4·5%; D, 1/37, 2·7%). With respect to IgM anti-prothrombin and IgG or IgM anti-β2-GP I, the prevalence of VTE was higher in both groups A and C than in group D, but no statistical difference in prevalence was found between groups A and C. Multivariate logistic regression analysis of risk factors for VTE confirmed that the co-existence of IgG anti-prothrombin and LA activity was the only significant risk factor for VTE (odds ratio, 19·13; 95% confidence intervals, 4·74–77·18).

Anti-phospholipid antibodies (aPL) are a heterogeneous group of antibodies that can be detected as anti-cardiolipin antibodies (aCL) and lupus anti-coagulant (LA) (Roubey, 1994). These antibodies are found frequently in the plasma of patients with systemic lupus erythematosus (SLE) (Cabiedes et al, 1995), and have been reported to be associated with clinical events such as arterial and/or venous thrombosis, thrombocytopenia and recurrent fetal loss in these patients (Ginsberg et al, 1995; D'Angelo et al, 1997; Galli et al, 1997a;Schulman et al, 1998; Galli & Barbui, 1999; Greaves, 1999).

It has been shown that the most common and characterized antigenic targets of aPL are β2-glycoprotein I (β2-GP I), recognized by aCL, and prothrombin, recognized by most LA (McNeil et al, 1990; Matsuura et al, 1992, 1994; Rao et al, 1996; Galli et al, 1997b). A number of clinical studies have established that anti-β2-GP I antibodies (anti-β2-GP I) are detectable in about 30% of patients with SLE, and that the presence of anti-β2-GP I constitutes a risk factor for arterial and/or venous thrombosis in patients with SLE (Koike & Matsuura, 1996; Barbui et al, 1997; Cucurull et al, 1999). On the other hand, some investigations showed that anti-prothrombin antibodies (anti-prothrombin) can be detected in about 50–60% of the patients with SLE, and that these antibodies correlate with venous thrombosis but not with arterial thrombosis (Forastiero et al, 1997; Palosuo et al, 1997; Munoz-Rodriguez et al, 2000). However, it has been reported that the LA assay correlates best with a history of both arterial and venous thrombosis, and that the enzyme-linked immunosorbent assay (ELISA) for anti-β2-GP I or anti-prothrombin does not give additional information for a thrombotic risk in SLE patients (Horbach et al, 1996).

Recently, it was shown that both anti-prothrombin and anti-β2-GP I could be divided to two types, one which shows LA activity and one that does not (Roubey, 1994; Galli & Barbui, 1999). It is not yet clear whether or not the existence of LA activity in these aPLs is related to the risk factors for VTE.

In the present study, we studied anti-prothrombin and anti-β2-GP I using a specific ELISA system in addition to LA activity detected using a phospholipid-dependent coagulation assay in 124 patients with SLE and found that the co-existence of IgG anti-prothrombin and LA activity correlated best with a history of VTE in patients with SLE.

Patients and methods

Patients We studied plasma samples from 124 patients (116 women, eight men; 17–71 years of age; mean, 41·6 years) with SLE. Diagnosis of SLE was made according to the revised criteria of the American Rheumatism Association. Clinical history revealed that venous thromboembolism (VTE) had occurred in 20 of the SLE patients: deep vein thrombosis (DVT; 12 cases) and pulmonary embolism (PE; eight cases). Diagnosis of DVT and PE was made on the basis of clinical manifestations and findings on duplex scanning, radioisotope venography, contrast venography and radioisotope lung scanning. For the control group, we studied 80 plasma samples from normal healthy volunteers (74 women, six men; 22–60 years of age; mean, 39·8 years) who worked at the Osaka University Hospital. None of them had any history of thrombotic complications, and there was no abnormality in the blood examinations (blood cell counts, coagulation tests, liver function tests and examinations for autoimmune). Blood samples were taken into vacuated tubes (5·0 ml total volume, Sekisui Medical, Osaka, Japan) containing 0·5 ml of 3·13% trisodium citrate (Na3C6H5O7·2H2O). Platelet-poor plasma was prepared by centrifuging at 2800 g for 15 min. The plasma sample was frozen at −80°C until batch assays were performed.

ELISA for aPLs Recent studies have indicated that anti-β2-glycoprotein I and anti-prothrombin do not recognize the native forms of β2-GP I and prothrombin on plain polystyrene ELISA plates, but do bind to the conformationally changed structures of β2-GP I and prothrombin coated on γ-irradiated polystyrene ELISA plates (Matsuura et al, 1994; Roubey, 1994; Igarashi et al, 1996; Koike & Matsuura, 1996; Galli & Barbui, 1998, 1999). In this study, we used a specific ELISA system for detecting the antibodies to prothrombin or β2-GP I, in which human prothrombin or β2-GP I (Diagnostica Stago, Asnieres, France) was immobilized directly on γ-irradiated polystyrene ELISA plates. Such plates (Nunc-Immunoplate, Maxi-Sorp, Kamstrup, Roskilde, Denmark) were coated overnight at 4°C with 50 µl per well of human prothrombin or β2-GP I suspended at a concentration of 10 µg/ml in Tris-buffered saline (TBS, 50 mmol/l Tris-HCl, 0·1 mol/l NaCl, pH 7·4). The wells were blocked for 60 min at room temperature with 50 µl of TBS containing 1·0% bovine serum albumin (BSA, Sigma, St Louis, MO, USA), and then washed three times with TBS containing 0·1% Tween 20 (TBS Tween). Thereafter, 50 µl of plasma sample (diluted 101 times with 1·0% BSA–TBS, 0·1% Tween 20) was added to each well. After 60 min of incubation at room temperature, the wells were washed with TBS–Tween. Horseradish peroxidase-conjugated goat anti-human IgG (µ-chain specific, A-2290) or IgM (µ-chain specific, A-4290) F(ab′)2 fragment of affinity isolated Ab (Sigma) was used, and the colour was developed by means of tetramethylbenzidine (TMB) solution (Moss Inc., MD, USA). The OD was measured at 450 nm. The OD of blank wells (i.e. coated only with TBS) was subtracted from the OD in the antigen-coated wells to account for non-specific binding. Monoclonal anti-human β2-GP I and anti-human prothrombin (MBL, Nagoya, Japan) were used in each assay as a positive control and selected normal plasma samples were used as a negative control.

Detection of lupus anti-coagulant (LA) activity The LA activity was detected using both the diluted Russell viper venom time (dRVVT) and STACLOT LA test. The dRVVT (Gradipore Ltd, Sydney, Australia) and STACLOT LA (Diagnostica Stago, Asnieres-sur-Seine, France) tests were performed using commercially available screening and confirmatory tests.

dRVVT-Gradipore Initially, 200 µl of plasma sample was incubated at 37°C. After 5 min of incubation, 200 µl of LA-screen reagent (simplified dRVVT reagent), which had been pre-incubated at 37°C, was added and the clotting time was then measured with an Amelung KC-10A (Heinrich Amelung GmbH, Lemgo, Germany). This time was designated as LA-screen. A 200 µl volume of plasma sample was incubated at 37 °C. After 5 min of incubation, 200 µl of LA-confirm reagent (phospholipid-rich reagent), which had been pre-incubated at 37°C, was added, and the clotting time was then measured with the Amelung KC-10A. This time was designated as LA-confirm. The ratio of the clotting time (LA-screen/LA-confirm) was calculated, and a ratio > 1·3 was defined as dRVVT-positive.

STACLOT LA Initially, 50 µl of plasma sample was incubated at 37°C with 50 µl of buffer in Tube 1. In Tube 2, a similar volume of sample was incubated with 50 µl of hexagonal phase phosphatidylethanolamine (HPE). After 9 min of incubation at 37°C, 50 µl of normal human platelet-poor plasma was added to each tube. After 1 min of incubation, 100 µl of activated partial thromboplastin time (APTT)-reagent was added to Tube 1 and Tube 2, and these were incubated for 5 min at 37°C. Then, 100 µl of 0·0025 mol/l CaCl2, which had been pre-incubated at 37°C, was added to Tube 1 and Tube 2. The clotting time for Tube 1 and Tube 2 was measured with the Amelung KC-10A, and when the clotting time in Tube 2 was greater (≥ 8 s) than that of Tube 1, LA was defined as positive.

Statistical analysis Fisher's exact probability test was used to evaluate the association between the prevalence of VTE and the presence or absence aPL. As an approximation of the relative risk, odds ratios (OR) were calculated for several putative risk factors using univariate logistic regression analysis with the statistical program stat flex (Artech, version 4·2, Osaka, Japan). An OR was considered to be statistically significant when the lower limit of the 95% confidence intervals (CI) was > 1·0. The variable that achieved statistical significance in the univariate logistic regression analysis was tested in a multivariate logistic regression analysis by the stat flex program.

Results

Prevalence of anti-prothrombin, anti-β2-GP I and LA activity

Here we studied the levels of anti-β2-GP I and anti-prothrombin Abs in 80 healthy control subjects. Each of the antibody levels detected using ELISA in the 80 control subjects were log transformed to approximate normality using the stat flex program before performing statistical analysis. The mean + 3SD of each antibody in 80 healthy control subjects was chosen as the cut-off point. The cut-off values (mOD) for anti-prothrombin and anti–β2-GP I were 500·8 (IgG anti-prothrombin) and 301·2 (IgM anti-prothrombin), 392·1 (IgG anti-β2-GP I), and 299·1 (IgM anti-β2-GP I). A result was regarded as positive when the absorbance exceeded each cut-off value. The prevalence of anti-prothrombin and anti-β2-GP I in the 124 patients with SLE was as follows: IgG anti-prothrombin, 65 cases (52·4%); IgM anti-prothrombin, 26 cases (21·0%); IgG anti-β2-GP I, 34 cases (27·1%); IgM anti-β2-GP I, 16 cases (12·9%). Using both dRVVT and STACLOT LA tests, we detected LA activity in 57 of the 124 patients with SLE.

Relationship between venous thromboembolism and anti-prothrombin or anti-β2-GP I with or without LA activity

The 124 patients with SLE were divided into four groups according to the results of ELISA and LA assay for each aPL: group A, ELISA+ and LA+ group B, ELISA+ and LA; group C, ELISA and LA+ group D, ELISA and LA.

Regarding the IgG anti-prothrombin, the prevalence of VTE was 45·7% (16/35 cases) in group A; 6·7% (2/30 cases) in group B; 4·5% (1/22 cases) in group C; and 2·7% (1/37 cases) in group D. The prevalence of VTE was by far the highest in group A (P < 0·001, Fisher's exact probability test) compared with that in the other groups (Fig 1). In those patients with the IgM anti-prothrombin, the prevalence of VTE was 25·0% (5/20 cases) in group A; 0% (0/6 cases) in group B; 32·4% (12/37 cases) in group C; and 4·9% (3/61 cases) in group D. The prevalence of VTE was higher in both groups A and C than in group D, but no statistical difference in prevalence was found between group A and group C (Fig 1).

Figure 1.

 Prevalence of venous thromboembolism (VTE) in SLE patients with or without anti-prothrombin and/or LA activity. The 124 patients with SLE were divided into four groups (A–D) according to the results of anti-prothrombin ELISA and LA assay. Statistical analysis was performed using Fisher's exact probability test.

With respect to IgG anti-β2-GP I, the prevalence of VTE was 39·1% (9/23 cases) in group A; 9·1% (1/11 cases) in group B; 23·5% (8/34 cases) in group C; and 3·6% (2/56 cases) in group D. In the case of the IgM isotype, the prevalence of VTE was 33·3% (3/9 cases) in group A; 0% (0/7 cases) in group B; 29·2% (14/48 cases) in group C; and 5·0% (3/60 cases) in group D. For both IgG and IgM isotypes, the prevalence of VTE was higher in both groups A and C than in group D, but no statistical difference in prevalence was found between group A and group C (Fig 2).

Figure 2.

 Prevalence of venous thromboembolism (VTE) in SLE patients with or without anti-β2-GP I and/or LA activity. The 124 patients with SLE were divided into four groups (A–D) according to the results of anti-β2-GP I ELISA and LA assay. Statistical analysis was performed using Fisher's exact probability test.

Univariate and multivariate logistic regression analysis

The univariate logistic regression analysis of IgG anti-prothrombin and IgG anti-β2-GP I with or without LA activity indicated that the significant risk factors for VTE were the co-existence of IgG anti-prothrombin and LA activity (OR, 17·89; 95% CI, 5·37–59·61) and that of IgG anti-β2-GP I and LA activity (OR, 5·26; 95% CI, 1·85–14·96; Table I). The presence of IgG anti-prothrombin, IgG anti-β2-GP I or LA activity alone was not a significant risk factor for VTE in SLE patients. However, when multivariate logistic regression analysis was performed, the co-existence of IgG anti-β2-GP I and LA activity was not a significant risk factor for VTE, but it was clear that the co-existence of IgG anti-prothrombin and LA activity was still a significant risk factor (OR, 19·13; 95% CI, 4·74–77·18). Concerning the analysis of IgM anti-prothrombin and IgM anti–β2-GP I, the univariate logistic regression analysis indicated that the co-existence of these IgM antibodies with LA activity was not a significant risk factor for VTE activity (IgM anti-prothrombin, OR, 1·98, 95% CI, 0·63–6·25; IgM anti–β2-GP I, OR, 2·88, 95% CI, 0·66–12·64; Table II).

Table I.   Univariate and multivariate logistic regression analysis of risk factors for venous thromboembolism (VTE).


ELISA


LA

Number
of cases
Prevalence of
VTE %
(number of cases)

Univariate analysis
OR (95%CI)

Multivariate analysis
OR (95%CI)
  • *

    Results shown to be statistically significant.

  • OR, odds ratio; 95%CI, 95% confidence intervals.

  • An OR was considered to be statistically significant when the lower limit of the 95%CI was > 1·0.

IgG anti-prothrombin
 ++3545·7 (16)17·89* (5·37–59·61)19·13* (4·74–77·18)
 +306·7 (2)0·30 (0·07–1·38) 
 –+224·5(1)0·21 (0·03–1·64) 
 –372·7 (1)0·10 (0·01–0·07) 
IgG anti-β2GP I
 ++2339·1 (9 )5·26* (1·85–14·96)0·88 (0·24–3·27)
 +119·1 (1)0·49 (0·06–4·10) 
 –+3423·5 (8)2·00 (0·74–5·43) 
 –563·6 (2)0·10 (0·02–0·47) 
Table II.   Univariate and multivariate logistic regression analysis of risk factors for venous thromboembolism (VTE).
ELISALANumber
of cases
Prevalence of
VTE %
(number of cases)
Univariate analysis
OR (95%CI)
Multivariate analysis
OR (95%CI)
  • *

    Results shown to be statistically significant.

  • OR, odds ratio; 95% CI, 95% confidence intervals.

  • An OR was considered to be statistically significant when the lower limit of the 95% CI was > 1·0.

IgM anti-prothrombin
 ++2025·0 (5)1·98 (0·63–6·25) 
 +60·0 (0 ) 
 –+3732·4 (12 )4·74* (1·74–12·90)2·30 (0·53–9·98)
 –614·9 (3 )0·14 (0·04–0·53) 
IgM anti-β2GP I
 ++933·3 (3)2·88 (0·66–12·64) 
 +7 0·0 (0 ) 
 –+4829·2 (14)4·80* (1·70–13·60)2·62 (0·57–11·98)
 –605·0% (3)0·14 (0·04–0·53) 

Combinations of anti-prothrombin, anti-β2-GP I, and LA activity in SLE patients with venous thromboembolism

Twenty (16·1%) of the 124 patients with SLE had VTE: 12 of these patients had DVT; and eight, PE (Table III). The prevalence of anti-prothrombin, anti-β2-GP I and LA activity in the 20 patients with VTE was 90% (18 cases), 50% (10 cases) and 85% (17 cases) respectively. Nine of these VTE patients had IgG anti-prothrombin, IgG anti-β2-GP I and LA activity; seven had IgG anti-prothrombin and LA activity but no IgG anti-β2-GP I; one had IgG anti-prothrombin and IgG anti-β2-GP I but no LA activity; one had IgG anti-prothrombin only; one had LA activity only; and one had neither anti-prothrombin and anti-β2-GP I nor LA activity.

Table III.   Clinical characteristics of 20 SLE patients with venous thromboembolism (VTE).
Patient
number
Age/sexComplicationsLAIgG
anti-prothrombin
IgM
anti-prothrombin
IgG
anti-β2GP I
IgM
anti-β2GP I
  1. The LA activity was detected by use of both the dRVVT and STACLOT LA test.

  2. The mean + 3SD of each antibody in 80 normal control subjects was decided as the normal cut-off point.

  3. The cut-off values for IgG anti-prothrombin, IgM anti-prothrombin, IgG anti-β2GP I, and IgM anti-β2GP I were 500·8 (mOD), 301·2, 392·1 and 299·1 respectively.

 141/FPE+5363171872313
 252/MDVT+6361664
 344/FPE+8285971415393
 451/FDVT+1574
 541/FPE+897– 
 626/FDVT+782840
 737/FDVT+7631280
 850/FDVT+729628
 945/FDVT+1247
1052/FDVT+5901528
1138/FDVT+854887
1239/FDVT+1092660594
1352/FPE+1815
1427/FDVT+1572927449716
1545/FPE+1465
1652/FPE+549
1771/FDVT1115
1848/FPE938828
1949/FPE+
2032/FDVT

Discussion

A number of previous studies have shown that anti-phospholipid antibodies (aPL) do not bind primarily to the negatively charged phospholipid itself but rather to complexes of the phospholipid and plasma proteins, and that the most common and best characterized antigenic targets are β2-glycoprotein I (β2-GP I) recognized by aCL and prothrombin recognized by most LA (McNeil et al, 1990; Matsuura et al, 1992, 1994; Rao et al, 1996 Galli et al, 1997b; Hanly & Smith, 2000; Lakos et al , 2000). The detection of anti-β2-glycoprotein I antibodies (anti-β2-GP I) and anti-prothrombin antibodies (anti-prothrombin) are of clinical importance, because they are known to be associated with arterial and/or venous thrombosis, thrombocytopenia, and fetal loss in patients with SLE (Roubey, 1994; Ginsberg et al, 1995; D'Angelo et al, 1997; Schulman et al, 1998).

In previous studies, we compared the prevalence of arterial thrombosis, venous thrombosis and thrombocytopenia in 180 SLE patients with or without β2-GP I-dependent aCL and/or LA activity, and found that the prevalence of arterial thrombosis was obviously higher in the SLE patients who had both aCL and LA activity than in the other SLE patients bearing aCL or LA activity alone or neither of them (Nojima et al, 1997). We also showed that the vast majority of patients with arterial thrombosis in the aCL and LA double-positive group had moderate thrombocytopenia (Nojima et al, 1998). Furthermore, when we studied the effects of aCL and/or LA on the enhancement of platelet activation induced by a low concentration of ADP, we demonstrated that platelet activation was significantly induced by patients' plasma or IgG fractions containing both aCL and LA activity but not by patients' plasma or IgG fractions containing either aCL or LA activity (Nojima et al, 1999). These findings raise the possibility that β2-GP I-dependent aCL and LA activity may co-operate to promote platelet activation, and may be involved in the pathogenesis of arterial thrombosis in patients with SLE.

On the other hand, a VTE, such as a DVT or PE, is also one of the common manifestations in SLE patients with aPL (Galli & Barbui, 1999). However, judging from our recent results, the co-existence of aCL and LA activity was not considered to be a risk factor for venous thrombosis, and so it was unclear whether aPL increases the risk of VTE.

Anti-prothrombin has been reported to correlate with a history of venous thrombosis (Palosuo et al, 1997; Galli & Barbui, 1999; Munoz-Rodriguez et al, 2000), whereas most clinical studies suggested that anti-prothrombin measured using ELISA was not reliable as a risk factor for venous thromboembolic events (D'Angelo et al, 1997; Galli & Barbui, 1998, 1999). Some studies suggest that IgG and IgM antibodies to the calcium-mediated prothrombin–phosphatidylserine complex were risk factors for thrombosis in patients suffering from SLE (Galli & Barbui, 1999). However, our preliminarily studies suggest the possibility that an ELISA using phosphatidylserine-coated plates may also detect antibodies that are reactive with some components of human plasma or bovine serum protein bound to the phosphatidylserine-coated plates. Therefore, we used a specific ELISA for detecting the anti-prothrombin antibodies, in which prothrombin was directly immobilized on γ-irradiated polystyrene plates.

Recently, it was reported that LA is the most significant risk factor for VTE (Horbach et al, 1996). As it has been suggested that LA activity can be caused by anti-prothrombin and/or anti-β2-GP I (Galli et al, 1995, 1998; Horbach et al, 1998), we studied the association of anti-prothrombin and anti-β2-GP I with or without LA activity in SLE patients with VTE.

We divided our 124 SLE patients into four groups according to the results of ELISA and LA assay for each aPL, as indicated in Patients and methods. It is important to note that the prevalence of VTE was quite high in the IgG anti-prothrombin ELISA+ LA+ group (45·7%, P < 0·001) compared with that in the IgG anti-prothrombin ELISA+ LA group (6·7%), the IgG anti-prothrombin ELISA LA+ group (4·5%) and the IgG anti-prothrombin ELISA LA group (2·7%). In contrast, with respect to IgM anti-prothrombin and IgG or IgM anti–β2-GP I, the prevalence of VTE was higher in both ELISA+ LA+ and ELISA LA+ groups than in the ELISA LA group, but no statistical difference in prevalence was found between the ELISA+ LA+ group and ELISA LA+ group.

The univariate logistic regression analysis indicated that the co-existence of IgG anti-prothrombin and LA activity was the most significant risk factors for VTE (OR, 17·89; 95% CI, 5·37–59·61), but that the co-existence of IgG anti-β2-GP I and LA activity also appeared to be significant risk factors for it (OR, 5·26; 95% CI, 1·85–14·96). However, when multivariate logistic regression analysis was performed, it became clear that the only significant risk factors for VTE was the co-existence of IgG anti-prothrombin and LA activity (OR, 19·13; 95% CI, 4·74–77·18). These findings suggest the possibility that the co-existence of IgG anti-prothrombin and LA activity may be one of the important factors in the pathogenesis of VTE in patients with SLE. However, the presence of anti-prothrombin, anti-β2-GP I or LA activity alone may not be sufficient for the pathogenesis of VTE in SLE patients.

Recently, it was suggested that LA activity can be caused by anti-prothrombin and/or anti-β2-GP I, and that the two antibodies may be discriminated by means of specific coagulation profiles of the Kaolin clotting time (KCT) and the dRVVT, both of which are well-known methods for detecting LA activity (Galli et al, 1995, 1998; Horbach et al, 1998). It was also suggested that the KCT profile reflects mainly the presence of anti-prothrombin, whereas the dRVVT profile is associated mostly with anti-β2-GP I, and that the LA activity measured by dRVVT might be associated with the prevalence of thrombosis compared with the LA activity measured using the KCT (Galli et al, 1995, 1998, 1999; Horbach et al, 1998).

However, in our 57 LA-positive patients; the LA activity was totally dependent on anti-prothrombin or on anti-β2-GP I in 13 and one case(s), respectively, whereas in the majority of these patients (43 cases), both anti-prothrombin and anti-β2-GP I simultaneously contributed to the LA activity. It is significant to note that the vast majority of SLE patients with VTE had both IgG anti-prothrombin and LA activity, and that half of the patients with VTE had no anti-β2-GP I, suggesting that an important risk factor for VTE is not the co-existence of IgG anti-β2-GP I and LA activity but the co-existence of IgG anti-prothrombin and LA activity. The multivariate logistic regression analysis of risk factors for VTE confirmed that the only significant risk factor for VTE was the co-existence of IgG anti-prothrombin and LA activity.

Recently, it was suggested that the complexes of phospholipid and plasma proteins such as protein C, protein S and annexin V are also recognized by aPL Abs, and these Abs may also be involved in thrombotic complications (Roubey, 1994; Galli & Barbui, 1999). Thus, we are now preparing to study the prevalence of antibodies against phospholipid-binding plasma proteins (β2-GP I, prothrombin, protein C, protein S and annexin V) detected using specific ELISA systems in our patients with SLE, and to investigate the relationships between these antibodies and the thrombotic and thrombocytopenic complications.

Ancillary