SEARCH

SEARCH BY CITATION

Keywords:

  • antiphospholipid syndrome;
  • beta2-glycoprotein I;
  • domain I;
  • obstetric complication;
  • thrombosis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Summary. Background: Diagnosis of the antiphospholipid syndrome (APS) is difficult as a result of limited specificity of existing assays for detecting clinically relevant antiphospholipid antibodies. Anti-beta2-glycoprotein I (beta2GPI) antibodies play a central role in the disease process of APS. Objectives: We have investigated the relation between antiphospholipid antibodies with specificity for domain I of beta2GPI and thrombosis/pregnancy morbidity in an international multicenter study. Patients/methods: Four hundred and seventy-seven patients derived from nine different centres met the inclusion criterion of having anti-beta2GPI antibodies in their plasma/serum. Clinical data and results of tests for lupus anticoagulant, anti-cardiolipin antibodies and anti-beta2GPI antibodies were established at the different centres of inclusion. After being re-tested for the presence of IgG and/or IgM anti-beta2GPI antibodies, the samples were tested for the presence of IgG-directed against domain I of beta2GPI and results were correlated with the thrombotic and obstetric history. Results: Re-testing for the presence of anti-beta2GPI antibodies resulted in inclusion of 442/477 patients. IgG class anti-domain I antibodies were present in plasma of 243/442 patients (55%). 201/243 (83%) had a history of thrombosis. This resulted in an odds ratio of 3.5 (2.3–5.4, 95% confidence interval) for thrombosis. Anti-domain I IgG antibodies were also significantly correlated with obstetric complications [odds ratio: 2.4 (1.4–4.3, 95% confidence interval)]. Conclusion: In this multicenter study, the detection of IgG antibodies that are directed against domain I of beta2GPI proved to be more strongly associated with thrombosis and obstetric complications than those detected using the standard anti-beta2GPI antibody assay.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

The antiphospholipid syndrome (APS) is a heterogeneous disease affecting many disciplines of medicine including hematology, immunology, rheumatology, obstetrics and neurology [1]. The diagnosis is clinically determined by a history of vascular thrombosis and/or specific obstetric complications [2]. To make the diagnosis of APS, antiphospholipid antibodies should be persistently present in the blood [3]. As thrombosis and obstetric complications have a high incidence, a large responsibility rests with the diagnostic laboratories because they determine whether a patient suffers from APS or not. Currently three assays are available for the diagnosis antiphospholipid syndrome according to the official Sydney Criteria: a phospholipid-dependent prolongation of clotting time [lupus anticoagulant (LAC)]; an anticardiolipin antibody (aCL) ELISA; and an anti-beta2-glycoprotein I (beta2GPI) antibody ELISA [4]. These assays detect overlapping but not necessarily identical subpopulations of antibodies. Frequently, persons without clinical signs which are compatible with APS test positive in one or more of these assays [5]. As a result of the high incidence of thrombosis and/or pregnancy morbidity in the general population, there is a strong need for highly specific assays to detect aPL in order to prevent over-diagnosis and unnecessary prolonged treatment.

In vitro and ex vivo experiments strongly suggest that antiphospholipid antibodies of the IgG isotype with affinity for plasma protein beta2-glycoprotein I are most relevant [6,7]. In this respect it was surprising that, in a meta-analysis of clinical studies on thrombosis, IgG and/or IgM autoantibodies directed against beta2-glycoprotein I did not show a significant association with thrombosis [8]. The domain-specificity of anti-beta2GPI antibodies might play a role in this. Several groups, among ourselves, have found evidence that a predominant part of the anti-beta2GPI IgG antibodies bind to domain I of beta2GPI [9–13]. In an attempt to develop more specific assays, we have recently discovered that IgG anti-beta2GPI antibodies with affinity for beta2-glycoprotein I can be divided into two groups: one group consisting of antiphospholipid antibodies that recognize domain I of beta2-glycoprotein I (type A antibodies) and a second group consisting of antiphospholipid antibodies with affinity for the other parts of beta2-glycoprotein I (type B antibodies) [12]. Some of us recently published in a single-centre study that antibodies against domain I of beta2-glycoprotein I correlated strongly with a history of thrombosis (odds ratio: 18.9, 95% confidence interval 6.8–52.3). In contrast, those with affinity for other domains were not significantly correlated with thrombosis (odds ratio: 1.1, 95% confidence interval 0.4–2.8) [13]. In the present study, we investigated the clinical significance of IgG anti-beta2GPI domain I antibodies in an international multicenter study.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Patients

Patients were included from nine different centres that are regarded to have specialized laboratories for detecting antiphospholipid antibodies (Fig. 1). Inclusion in this study was based on the presence of anti-beta2-glycoprotein I antibodies in the plasma of patients detected at least twice at least 12 weeks apart. Plasma/serum of 477 patients was sent to Sanquin on the basis of these inclusion criteria. All samples were re-tested at the Sanquin Blood Foundation to ensure homogeneity of anti-beta2GPI antibody testing. The in-house anti-beta2GPI ELISA is described in detail below. Four hundred and forty-two of the 477 plasma samples also tested positive in the in-house anti-beta2GPI antibody assay from Sanquin. Out of these 442 patients, 364 patients (82%) met the clinical Sydney Criteria for the diagnosis of the APS [4].

image

Figure 1.  Distribution of obtained optical densities (ODs) and domain ratios for different groups of patients. Figure displays the mean OD/ratio ±SEM. Black bars represent the mean OD/ratio of plasma of patients that have a history of one of the clinical symptoms. Differences between groups were calculated using Students t-test. *P < 0.05.

Download figure to PowerPoint

Primary and secondary APS was defined by meeting the Sydney criteria in the absence or presence of systemic lupus erythematosus [SLE; meeting at least four criteria from the American College of Rheumatology (ACR) for the classification of SLE] lupus-like disease (LLD; defined as meeting one to three ACR criteria) [14].

Clinical symptoms were registered according to the Sydney criteria for diagnosing APS [4]. The number of objectively verified arterial and venous thromboembolic events as well as the obstetric history in women was recorded by chart review. Computed tomographic scanning/magnetic resonance imaging was used for diagnosis of intracerebral vessel thrombosis. Typical electrocardiographic changes and elevated troponin I or T levels diagnosed myocardial infarction. During surgery, arteriography or thrombectomy diagnosed peripheral arterial thrombosis and thrombosis of the distal aorta. Retinal thrombosis was documented by fundus examination and fluorescence angiography. Ultrasonography/venography diagnosed a deep vein thrombosis, radionuclide lung scanning and CT-angiography diagnosed a pulmonary embolism and angiography was used for the diagnosis of portal vein thrombosis. Informed consent was obtained from all patients and the local ethics committee at each study center approved the study.

Design of the study

Invitations were sent to members of the European Forum on Antiphospholipid Antibodies to send plasma samples from patients positive for anti-beta2GPI antibodies to the Sanquin Blood Supply Centre, Amsterdam, the Netherlands. On arrival, all samples were retested on the presence of anti-beta2GPI IgG/IgM antibodies (in-house assay) to ensure homogeneity. Subsequently the samples were tested for IgG antibodies directed against domain I of beta2-glycoprotein I. At the time of testing for anti-domain I antibodies, no clinical and serological data (except anti-beta2GPI antibodies) were known to the Sanquin Blood Supply Centre. The presence of anti-domain I antibodies was determined at the Sanquin Blood Supply Centre. The clinical data were sent to Sanquin after all laboratory parameters were determined. Subsequently, associations between laboratory and clinical data were calculated at the Sanquin Blood Supply Centre using spss (SPSS Inc., Chicago, IL, USA). Chi-square statistics was applied to the data to compare the prevalence of thrombosis with serologic findings. Binary logistic regression was used to calculate odds ratios and 95% confidence intervals. Student’s t-test was used to calculate differences between the two groups.

Anti-beta2GPI IgG/IgM ELISA

To detect antibodies directed against the complete protein beta2GPI an in-house assay was used. Hydrophilic plates were coated with 5 μg mL−1 plasma purified beta2GPI for at least 12 h at 4 °C. Subsequently, the plates were blocked with 0.1% Tween/phosphate-buffered saline (PBS) followed by washing the plates three times with 0.1% Tween/PBS. Then the samples (duplicates) were diluted 1/100 in 0.1% Tween/PBS and added to the plate together with the blank, controls and calibration curves (RELARES, Leiden, the Netherlands). After 1 h, the plates were washed and incubated with either a polyclonal PO-labelled anti-human IgG or IgM antibody. Coloring was performed with 3′,5,5′-tetramethylbenzidine (TMB). The intra-assay CV of this assay, calculated from 20 individual samples on the same ELISA plate, was 8.9% (IgG) and 4.2 (IgM). The inter-assay CV, calculated from 19 individual samples performed on 10 different days, was 18% for IgG and 16% for IgM. In addition, one IgG and IgM sample was diluted to 12 different concentrations and each dilution was added five times to the plate to determine the linearity of the assay. For IgG we found a correlation coefficient of 0.998 (standard deviation: 5.9, VC: 11.0%) and for IgM we found a correlation coefficient of 0.996 (standard deviation: 5.9, VC: 7%). The cut-off of the assay was determined from the mean of 100 healthy individuals plus 10 times standard deviation.

Anti-domain I IgG ELISA

For the detection of antibodies directed against domain I of beta2GPI (specifically against the epitope G40-R43), we used a protocol that was previously published [13]. In short, domain I (3 μg mL−1) was coated on a hydrophobic and a hydrophilic plate. By coating domain I on a hydrophobic plate, the hydrophilic epitope R39-R43 will be available for binding antibodies. When domain I is coated on a negatively charged hydrophilic plate, the epitope R39-R43 will be involved in the binding to the plate and not available for binding of antibodies. Both types of plates were chosen in order to have an equal density of domain I coated to the surface which was checked with a monoclonal mouse anti-domain I antibody (mAb 3B7) that does not recognize epitope Arginine-39 to arginine-43. The plates were blocked with 3% bovine serum albumin/0.1% Tween/150 mM NaCL, pH 7.4 for 1 h. Then patient samples (duplicates) were diluted 1/100 in blocking solution was added and incubated for 1 h, followed by incubation of an anti-IgG peroxidase-labeled antibody to detect bound antibodies. TMB was used for staining and absorbance was read at 450 nm. To determine the presence of anti-domain I antibodies, the obtained optical density (OD) from a sample on a hydrophobic plate (epitope G40-R43 available for binding) was divided by the OD obtained from the hydrophilic plate (epitope G40-R43 unavailable for binding). A ratio greater than two discriminates between anti-beta2GPI antibodies that recognize epitope G40-R43 and anti-beta2GPI that recognize other parts of beta2GPI [13]. A sample that was regarded negative {OD did not reach the cut-of value [OD−(blank + three times SD)]} on the hydrophobic plate was given a zero. A sample that was regarded negative [OD−(blank + three times SD)] on the hydrophilic plate but positive on the hydrophobic plate was given > 2. A sample that was regarded negative on both plates was given zero. Cut-off values were determined from 50 plasma samples derived from healthy volunteers (mean plus three times SD). In all runs, a blank, negative control (normal pool plasma) and positive controls were included to calculate the cut-off. For the detection of anti-domain I IgG antibodies in plasma, the assay has an intra-assay CV of 12.65% (10 individual samples in the same test on the same day) and an inter-assay CV of 14.82 (eight independent samples on eight different days).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Laboratory and clinical characteristics are summarized in Table 1. Three hundred and sixty-four patients met the Sydney criteria meaning that they had either thrombosis or pregnancy complications as defined in the Sydney Criteria. All patients were tested for IgG antibodies with specificity for domain I. In blood of 243/442 (55%) patients anti-domain I IgG antibodies were detected (Table 1). Among them, 83% had a history of thrombosis (58% had a history of venous thrombosis and 41% had a history of arterial thrombosis). In the population of patients without anti-domain I antibodies (n = 199), 58% had a history of thrombosis. Of these, 41% had a history of venous thrombosis and 25% had a history of arterial thrombosis. The association between anti-domain I IgG antibodies and thrombosis resulted in an odds ratio of 3.5 (2.3–5.4, 95% confidence interval, Table 2).

Table 1.   Clinical and serological characteristics
 Total populationAnti-domain I IgGNo anti-domain I IgG
  1. Patients were selected on the presence of either anti-beta2GPI IgG antibodies or anti-beta2GPI IgM antibodies and subsequently tested for the presence of anti-domain I IgG antibodies. All laboratory and clinical variables were detected according to the official Sydney guidelines [4].

General characteristics
 Patients (n)442243199
 Female (n)352 (80%)188 (77%)164 (82%)
 Age [year, mean (range)] 43 (16–94) 40 (16–94) 47 (18–92)
Clinical characteristics
 APS-related clinical symptoms364 (82%)218 (90%)146 (73%)
 Thrombosis316 (71%)201 (83%)115 (58%)
  Arterial thrombosis149 (34%)100 (41%) 49 (25%)
  Venous thrombosis222 (50%)141 (58%) 81 (41%)
  Arterial and venous thrombosis 55 (12%) 40 (16%) 15 (8%)
 Pregnancy morbidity115/201 (57%) 69/102 (68%) 46/99 (46%)
Serological characteristics
 Antiphospholipid antibodies442 (100%)243 (100%)199 (100%)
  LAC279 (63%)177 (73%)102 (51%)
  aCL385 (87%)224 (92%)161 (81%)
  Anti-beta2GPI IgG420 (95%)239 (98%)181 (91%)
  Anti-beta2GPI IgM350 (79%)192 (79%)158 (79%)
  Anti-beta2GPI IgG and IgM312 (71%)192 (79%)120 (60%)
Diagnostic characteristics
 APS364 (82%)218 (90%)146 (73%)
  PAPS236 (53%)164 (67%) 72 (36%)
  Secondary APS128 (29%) 54 (22%) 74 (37%)
 SLE 93 (21%) 36 (15%) 57 (29%)
 LLD 35 (8%) 18 (7%) 17 (9%)
Table 2.   Association between aPL and thrombosis
 Odds ratio (95% confidence interval)
  1. To estimate whether there is a significant increase in association of anti-domain I IgG antibodies with thrombosis an odds ratio was calculated within the total population of 511 patients. *One is not included in 95% confidence interval. Bold: Significant association of assay with clinical symptom.

Anti-domain I IgG3.5 (2.3–5.4)*
Non-domain I Anti-beta2GPI IgG0.4 (0.3–0.6)
Anti-beta2GPI IgM0.9 (0.6–1.3)
LAC1.8 (1.1–3.1)*
aCL1.1 (0.6–2.1)

It is thought that antiphospholipid antibodies of the IgG type are better correlated with thrombosis than IgM antiphospholipid antibodies. Therefore, the association of anti-domain I IgG antibodies with thrombosis was re-tested in a subpopulation of 420 patients all positive for anti-beta2GPI IgG antibodies. Within this population of 420 patients, 239 patients were positive for anti-domain I IgG antibodies of which 198 patients had a history of thrombosis. This resulted in an odds ratio of 3.3 (2.1–5.2, 95% confidence interval).

As obstetric complications are another clinical criterion for APS, a possible association with anti-domain I IgG antibodies was also investigated. From 201 women that had been pregnant in the past a complete obstetric history was known (Table 3a). 115/201 women suffered from obstetric complication as described in the Sydney Criteria [4]. Anti-domain I antibodies were present in 102/201 patients of which 69 had suffered from obstetric complications in the past resulting in an odds ratio of 2.4 (1.4–4.3, Table 3b). In detail, we found a significant association with two out of the three obstetric criteria for APS; (i) history of fetal death beyond the 10th week of gestation [OR: 2.1 (1.2–3.7)]; (ii) history of premature birth(s) before the 34th week of gestation as a result of preeclampsia or placental insufficiency [OR: 2.0 (1.0–4.0)]. For anti-beta2GPI IgG or IgM antibodies or LAC we did not find an association with any of the obstetric criteria. For anticardiolipin antibodies we found a significant association with a history of obstetric complications (OR: 3.3, 1.7–6.6, 95% confidence interval), which was mostly as a result of the association with fetal death beyond the 10th week of gestation (OR: 3.0, 1.4–6.4, 95% confidence interval). As already suggested from the results presented in Tables 2 and 3a,b, differences in ODs or domain I ratio (= OD hydrophobic plate/OD hydrophilic plate) appeared to be larger in the thrombosis groups than in the pregnancy groups (Fig. 1). Recently, it was shown that the combination of LAC and anti-beta2GPI antibodies displayed a better association with thrombosis than the two parameters separately [6]. Therefore, we investigated the association between thrombosis and aCL and LAC in 442 patients all positive for anti-beta2GPI antibodies. For aCL and LAC we found an odds ratio of 1.1 (0.6–2.1, 95% confidence interval) and 1.8 (1.1–3.1, 95% confidence interval), respectively, indicating that only the combination of LAC and anti-beta2GPI antibodies increases its association with thrombosis. The combination of a positive LAC and positive aCL did not improve the association compared with positive for LAC alone.

Table 3.   Association between (a) pregnancy loss and aPL and (b) obstetric complications and aPL
 Obstetric complicationsHistory of three or more consecutive unexplained losses < 10 weeks gestationHistory of fetal death after 10 weeks before gestationHistory of premature birth(s) before 34 weeks due to preeclampsia or placental insufficiency
(a)
Women who had been pregnant Total (n = 201)115 (57%)32 (16%)81 (40%)39 (19%)
Anti-domain I IgG (n = 102)69 (68%)16 (16%)50 (49%)25 (25%)
LAC (n = 153)86 (56%)25 (16%)57 (37%)32 (21%)
aCL (n = 155)99 (64%)28 (18%)71 (46%)31 (20%)
Anti-beta2GPI IgG (n = 178)104 (58%)31 (17%)72 (40%)34 (19%)
Anti-beta2GPI IgM (n = 96)58 (60%)18 (19%)42 (44%)15 (16%)
  1. For the total population of 511 patients 201 patients had been pregnant in the past and a complete obstetric history was known. The obstetric complication were subdivided into three categories as described by the official criteria for APS (4); (a) a history of three or more consecutive unexplained losses < 10 weeks gestation; (b) a history of fetal death after 10 weeks gestation; (c) a history of one or more premature births before 34 weeks as a result of preeclampsia or placental insufficiency. To estimate whether there is a significant increase in association of anti-domain I IgG antibodies with obstetric complications an odds ratio was calculated in the population of 201 women with a known obstetric history. *One is not included in 95% confidence interval. Bold, significant association of assay with clinical symptom.

  Odds ratio for obstetric complications (95% confidence interval)Odds ratio for history of three or more consecutive unexplained losses < 10 weeks gestation (95% confidence interval)Odds ratio for history of fetal death after 10 weeks before gestation (95% confidence interval)Odds ratio for history of premature birth(s) before 34 weeks due to preeclampsia or placental insufficiency (95% confidence interval)
(b)
Anti-domain I IgG2.4 (1.4–4.3)*1.0 (0.5–2.1)2.1 (1.2–3.7)*2.0 (1.0–4.0)*
Anti-beta2GPI IgG1.5 (0.6–3.7)4.6 (0.6–35.7)1.1 (0.4–2.6)0.9 (0.3–2.5)
Anti-beta2GPI IgM1.3 (0.7–2.3)1.5 (0.7–3.2)1.3 (0.7–2.3)0.6 (0.3–1.3)
LAC0.9 (0.5–1.7)1.1 (0.4–2.8)0.6 (0.3–1.2)1.8 (0.7–4.7)
aCL3.3 (1.7–6.6)*2.3 (0.8–7.0)3.0 (1.4–6.4)*1.2 (0.5–2.7)

In the Sydney Criteria for the diagnosis of the APS, one of the recommendations is not to test the presence of antiphospholipid antibodies in elderly as this might lead to a high number of false-positive patients [4]. We have investigated the age distribution of patients positive for anti-domain I IgG antibodies compared with the rest of the patient population. For the population of patients with anti-domain I IgG antibodies, a mean age of 40 years (range: 16–94 years) was found compared with a mean age of 47 years (range: 18–92 years) for the population of patients without anti-domain I antibodies (P < 0.001, Table 4). As age is a risk factor for thrombosis, we investigated a possible difference in age between patients with anti-domain I IgG antibodies and thrombosis and patients with anti-domain I IgG antibodies without thrombosis. We found no significant difference [41 vs. 38 years (mean age), P = ns] between the two groups. But in the population of patients without anti-domain I IgG antibodies a significant difference in age was found between patients with thrombosis (mean age 49 years) and patients without thrombosis (mean age 43 years, P < 0.001).

Table 4.   Age differences between groups of patients with different aPL specificity
 Age (years, mean, range)Age differences between groups
  1. Differences in age between the different subgroups were calculated by log-transformation to ensure a Gaussian distribution followed by a Student’s t-test.

Group A: total population (n = 442)43 (16–94)Group A vs. B: P = 0.01 Group A vs. C: P < 0.01 Group B vs. C: P < 0.0001
Group B: anti-domain I IgG (n = 243)40 (16–94)
Group C: no anti-domain I IgG (n = 199)47 (18–92)
Group D: anti-domain I IgG and thrombosis (n = 201)41 (16–94)Group D vs. E: P = 0.20
Group E: anti-domain I IgG without thrombosis (n = 42)38 (19–73)
Group F: no anti-domain I IgG and thrombosis (n = 115)49 (20–92)Group F vs. G: P < 0.001
Group G: no anti-domain I IgG without thrombosis (n = 84)43 (18–83)

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

As a result of the high prevalence of thrombosis and/or obstetric complications in the general population, the diagnosis of the APS totally depends on serological markers. In several meta-analyses it was shown that none of the existing laboratory assays was specific enough for a definite diagnosis of APS [8,15]. In a recent single-centre study it was found that IgG antibodies with affinity towards the N-terminal part of beta2-glycoprotein I (domain I) were highly associated with thrombosis whereas antibodies against other domains of beta2GPI were not [13]. In the current study, we have validated this finding by conducting an international multi-center study including patients positive for anti-beta2GPI antibodies. In this population, 83% of the patients with domain I IgG antibodies had a history of vascular thrombosis compared with 58% of the patients without anti-domain I IgG antibodies (Table 1). These results are comparable with the first single center study in which we found that 83% of the patients with anti-domain I IgG antibodies had a history of thrombosis [8].

In the present study, we additionally analyzed the association of APS-defined obstetric complications in association with anti-domain I IgG antibodies. As with thrombosis, anti-domain I antibodies proved to be significantly correlated with obstetric complications, preferentially after 10 weeks of gestation. This association was calculated within a population of 201 women that had been pregnant and of which a complete obstetric history was known. As a result of limited data on APS-defined obstetric complications we cannot rule out some sort of selection bias. The fact that anti-domain I antibodies are not associated with a history of three or more consecutive unexplained losses < 10 weeks gestation might indicate different pathogenic mechanisms for the three APS-defined obstetric complications.

At present LAC is accepted to be the most specific test for thrombosis but the execution of LAC assays is complicated and sensitive to/influenced by many pre-analytical variables [16]. Moreover, LAC can be caused by anti-beta2GPI antibodies and by anti-prothrombin antibodies. Recently, it was shown that only LAC caused by anti-beta2GPI antibodies is related with thrombosis [6]. Unfortunately, the assay to discriminate between LAC caused by the two different antibodies is difficult to perform. This problem could be overcome by performing the anti-domain I IgG antibody ELISA as these antibodies are mainly responsible for LAC activity [13]. The anti-domain I IgG assay is easy to perform as it is in an ELISA format. Although this study was performed in a patient cohort selected for presence of anti-beta2GPI antibodies, we show again that a combination of a positive LAC and a positive anti-beta2GPI antibody ELISA correlates better with thrombosis than a positive anti-beta2glycoprotein I ELISA alone [OR LAC: 1.8 (1.1–3.1, 95% confidence interval)]. However, anti-domain I IgG antibodies showed an even better association with thrombosis [OR: 3.5 (2.3–5.4, 95% confidence interval)]. This might indicate that the anti-domain I IgG ELISA is more specific for thrombosis than the existing assays currently available. On the other hand, one should realize that the patient population was selected on positivity in the anti-beta2GPI IgG/IgM antibody ELISA. Therefore, conclusions related to antiphospholipid antibodies in general should be made with caution.

In general, IgG antiphospholipid antibodies are thought to be better correlated with thrombosis than IgM antibodies. As our selection criteria was positivity in the anti-beta2GPI IgG and/or IgM ELISA, the odds ratio of 3.5 for thrombosis of anti-domain I IgG antibodies could be biased as we only detected IgG anti-domain I antibodies. But the odds ratio of 3.5 almost remained the same [OR: 3.3 (2.1–5.2, 95% confidence interval)] when we calculated the association with thrombosis in the population of patients positive for anti-beta2GPI IgG antibodies. We did not detect anti-domain I IgM antibodies because we did not observe an increase in association with thrombosis in our previous study for anti-domain I IgM antibodies compared with anti-beta2GPI IgM antibodies directed against other parts of beta2GPI [8]. This was because of the fact that every plasma containing anti-beta2GPI IgM antibodies also contained anti-domain I IgM antibodies, making discrimination between negative and positive samples on the basis of our assay impossible.

Although all patients included in this study tested positive in the anti-beta2GPI ELISA, not all patients fulfiled classification criteria for APS. Apart from thrombosis and pregnancy morbidity indications for testing for aPL and hence discovery of ant-beta 2 antibodies had been the presence of SLE, LLD, livedo reticularis, transient ischemic attack or a family member with the antiphospholipid syndrome. In addition, this was a relative small group. Furthermore, the nine participating centers predominantly were reference centers for their country indicating a pre-selection of patients. This pre-selection together with small group of included persons without any clinical symptom might lead to an underestimation of the diagnostic value of detecting anti-domain I IgG antibodies.

One of the recommendations added to criteria to make the diagnosis of APS was to be reluctant in testing elderly for the presence of antiphospholipid antibodies [4]. It has been described that the prevalence of clinically non-significant antiphospholipid antibodies increases with age [17–19]. In this study, we found that patients with anti-domain I IgG antibodies were significantly younger than patients with antibodies directed against other parts of beta2GPI. As anti-domain I IgG antibodies are better correlated with thrombosis this might indicate that the prevalence of clinically non-significant anti-beta2GPI antibodies (with specificity against the other domains of beta2GPI) increases with age. Further evidence was found in the fact that patients with thrombosis but without anti-domain I IgG antibodies were older than patients without thrombosis (Table 4). This difference was not found in the group of patients that were positive for anti-domain I IgG antibodies (Table 4).

In conclusion, in this international multicenter study we could confirm that anti-domain I IgG antibodies are highly associated with thrombosis. But additional prospective studies should be undertaken before the anti-domain I ELISA can be included into the official criteria for the diagnosis APS.

Addendum

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Author contributions: conception and design: B. de Laat and P. G. de Groot. Analysis and interpretation of the data: B. de Laat, K. Mertens and P.G. de Groot. Including patients, analysis, critically reviewing the manuscript before submission: V. Pengo, I. Pabinger, J. Musial, A.E. Voskuyl, T. Kveder, I.E.M. Bultink, A. Ruffatti, B. Rozman, P. de Moerloose, F. Boehlen, J. Rand and Z. Ulcova-Gallova.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

For technical, logistical and analytical assistance the authors would like to thank J. van Keulen, R. Boertjes and M. Boon-Spijker from Sanquin (Sanquin, Amsterdam, the Netherlands), S. Koder (Medical University Vienna, Vienna, Austria), T Iwaniec (Jagiellonian University Medical College, Krakow, Poland), M. Tonello (University of Padua, Padua, Italy) and M. Raoufi (Montefiori Medical Center, New York, NY, USA). Grant support: Z. Ulcova-Gallova was funded by an unrestricted grant from the Czech Ministry of Education (MSM 0021620812). BdL was funded by the Netherlands heart Foundation (Grant 2006T053). Both funding sources played no role in the design, conduct or reporting of this study.

Disclosure of Conflict of Interests

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

The authors state that they have no conflict of interest.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References
  • 1
    Cervera R, Piette JC, Font J, Khamashta MA, Shoenfeld Y, Camps MT, Jacobsen S, Lakos G, Tincani A, Kontopoulou-Griva I, Galeazzi M, Meroni PL, Derksen RH, De Groot PG, Gromnica-Ihle E, Baleva M, Mosca M, Bombardieri S, Houssiau F, Gris JC, et al. Antiphospholipid syndrome: clinical and immunologic manifestations and patterns of disease expression in a cohort of 1000 patients. Arthritis Rheum 2002; 46: 101927.
  • 2
    Wong RC, Favaloro EJ. Clinical features, diagnosis, and management of the antiphospholipid syndrome. Semin Thromb Hemost 2008; 34: 295304.
  • 3
    De Laat B, Mertens K, De Groot PG. Mechanisms of disease: antiphospholipid antibodies-from clinical association to pathologic mechanism. Nat Clin Pract Rheumatol 2008; 4: 1929.
  • 4
    Miyakis S, Lockshin MD, Atsumi T, Branch DW, Brey RL, Cervera R, Derksen RH, 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). J Thromb Haemost 2006; 4: 295306.
  • 5
    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: 399401.
  • 6
    De Laat HB, Derksen RH, Urbanus RT, Roest M, De Groot PG. Beta2-glycoprotein I-dependent lupus anticoagulant highly correlates with thrombosis in the antiphospholipid syndrome. Blood 2004; 104: 3598602.
  • 7
    Pengo V, Biasiolo A, Pegoraro C, Cucchini U, Noventa F, Iliceto S. Antibody profiles for the diagnosis of antiphospholipid syndrome. Thromb Haemost 2005; 93: 114752.
  • 8
    Galli M, Luciani D, Bertolini G, Barbui T. Anti-beta 2-glycoprotein I, antiprothrombin antibodies, and the risk of thrombosis in the antiphospholipid syndrome. Blood 2003; 102: 271723.
  • 9
    Ioannou Y, Pericleous C, Giles I, Latchman DS, Isenberg DA, Rahman A. Binding of antiphospholipid antibodies to discontinuous epitopes on domain I of human beta(2)-glycoprotein I: mutation studies including residues R39 to R43. Arthritis Rheum 2007; 56: 28090.
  • 10
    McNeeley PA, Dlott JS, Furie RA, Jack RM, Ortel TL, Triplett DA, Victoria EJ, Linnik MD. Beta2-glycoprotein I-dependent anticardiolipin antibodies preferentially bind the amino terminal domain of beta2-glycoprotein I. Thromb Haemost 2001; 86: 5905.
  • 11
    Yu P, Passam FH, Yu DM, Denyer G, Krilis SA. Beta2-glycoprotein I inhibits vascular endothelial growth factor and basic fibroblast growth factor induced angiogenesis through its amino terminal domain. J Thromb Haemost 2008; 6: 121523.
  • 12
    Iverson GM, Victoria EJ, Marquis DM. Anti-beta2 glycoprotein I (beta2GPI) autoantibodies recognize an epitope on the first domain of beta2GPI. Proc Natl Acad Sci USA 1998; 95: 155426.
  • 13
    De Laat B, Derksen RH, Urbanus RT, De Groot PG. IgG antibodies that recognize epitope Gly40-Arg43 in domain I of beta 2-glycoprotein I cause LAC, and their presence correlates strongly with thrombosis. Blood 2005; 105: 15405.
  • 14
    Tan EM, Cohen AS, Fries JF, Masi AT, McShane DJ, Rothfield NF, Schaller JG, Talal N, Winchester RJ. The 1982 revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1982; 25: 12717.
  • 15
    Galli M, Luciani D, Bertolini G, Barbui T. Lupus anticoagulants are stronger risk factors for thrombosis than anticardiolipin antibodies in the antiphospholipid syndrome: a systematic review of the literature. Blood 2003; 101: 182732.
  • 16
    Tripodi A. Laboratory testing for lupus anticoagulants: a review of issues affecting results. Clin Chem 2007; 53: 162935.
  • 17
    Meroni PL, Mari D, Monti D, Coppola R, Capri M, Salvioli S, Tincani A, Gerli R, Franceschi C. Anti-beta 2 glycoprotein I antibodies in centenarians. Exp Gerontol 2004; 39: 145965.
  • 18
    Ramos-Casals M, García-Carrasco M, Brito MP, López-Soto A, Font J. Autoimmunity and geriatrics: clinical significance of autoimmune manifestations in the elderly. Lupus 2003; 12: 34155.
  • 19
    Andersen-Ranberg K, Hoier-Madsen M, Wiik A, Jeune B, Hegedus L. High prevalence of autoantibodies among Danish centenarians. Clin Exp Immunol 2004; 138: 15863.