Antiphospholipid Antibodies and the Antiphospholipid Syndrome I: Pathogenesis, Clinical Features, Diagnosis, and Management

Welcome to the first of two issues of Seminars in Thrombosis and Hemostasis on the topic of antiphospholipid (aPL) antibodies and the antiphospholipid syndrome (APS). The next issue will deal with the limitations, standardization, and clinical utility of laboratory testing for aPL antibodies and APS. The current issue concerns the pathogenesis, clinical features, diagnosis, and management of APS.

The opening article is by Harris and Pierangeli, who provide a concise review of the terminology of APS. They also cover many of the major milestones in the understanding of this syndrome by arbitrarily defining three time periods, beginning with an “observational period” from 1953 to 1983, when the relatively infrequent observations were largely based on the lupus anticoagulant test. This was followed by a period of “exponential growth in interest” from 1983 to 1995, triggered by the publication in 1983 of a specific anticardiolipin test using a radioimmunoassay method.[1] The method for performing the anticardiolipin test was converted to a more easily performed enzyme-linked immunosorbent assay (ELISA) during this period, and the term antiphospholipid syndrome was introduced on the basis that both lupus anticoagulants and anticardiolipin antibodies were part of a larger group of “antiphospholipid” antibodies. Subsequently, the terminology primary/secondary APS and associated concepts were also introduced. In essence, this nomenclature was developed by rheumatologists who were used to a primary/secondary terminology to subgroup patients with Sjögren's syndrome and thus applied a similar terminology to patients with APS. The catastrophic APS terminology was also developed roughly midway through this crucial 12-year period. A third period of “consolidation and refinement” can be identified from 1995 to the current day and includes the development of the 1999 Sapporo[2] and the subsequently revised 2006 Sydney[3] classification criteria for APS. The use of the Sapporo and Sydney criteria have led to the formation of APS patient registries both in North America and Europe, crucial for conducting prospective studies, in contrast with the retrospective studies that the first two periods of knowledge-building regarding APS were largely based on. The authors remind us that the rationale and driving force for the development and refinement of these criteria was to produce a sensitive and, more importantly, specific definition of APS to enable more rigorous studies to be performed using cohorts of patients who “unquestionably” had the syndrome. For this reason, these criteria do not enable a classification of APS in patients with only so-called aPL associated features (e.g., cardiac valve abnormalities, livedo reticularis, nephropathy, and thrombocytopenia), where evidence of an association with APS exists but is not conclusive. The authors also emphasize a frequently overlooked point in that these criteria are classification criteria whose primary aim is to correctly classify patients (as their name implies) for inclusion in clinical research studies, rather than being primarily diagnostic criteria. As such, they point out that there are patients with APS who would not fulfill the Sapporo or Sydney criteria because their manifestations are primarily nonthrombotic, or their thrombotic manifestations have not been documented as vigorously, or their number of pregnancy-related events is less than that required by the criteria. However, whereas these patients cannot be included in prospective studies of APS patients, they should still be managed appropriately as APS patients, including anticoagulation as required clinically.

Harris and Pierangeli also briefly mention issues related to the anticardiolipin antibody and anti–β2 glycoprotein I (anti-β2GPI) antibody assays. Although these issues are covered in much greater detail within the subsequent issue of Seminars in Thrombosis and Hemostasis,[4] the authors contend that reproducible quantitative (and standardized) anticardiolipin assays may never be achieved. Using the classification criteria–defined patient registries to evaluate the sensitivity and specificity of the “newer” tests for APS including anti-β2GPI may be helpful, but one has to consider the significant but underrecognized effect of incorporation bias, which favors the particular anticardiolipin (aCL), lupus anticoagulant (LA), and anti-β2GPI assays that were initially used to fulfill the laboratory criteria.

In the second article of this issue of Seminars in Thrombosis and Hemostasis, Ronald Asherson gives a personal historical account of how the various subgroups of the aPL antibody syndrome came to be recognized and published. His article includes some potentially controversial comments regarding the initial recognition of the entity that is now known as primary APS (PAPS). He contends that his publication of an original description of primary APS prepared in 1985 was delayed until 1989 because of a prevailing view that such patients also had concomitant systemic lupus erythematosus (SLE), which history has proved to be incorrect. The concept of PAPS has proved to be a crucial one as most patients with APS are now considered to have this entity without any evident concomitant autoimmune disease. Asherson also discusses the primary plus syndrome, which describes patients with features of APS who have one or two clinical features (such as Raynaud's phenomenon and certain forms of renal disease) that are not included in the current classification criteria for SLE or other connective tissue diseases. Occasionally, these patients will also have positive laboratory tests for rheumatoid factor, cryoglobulins, or immune complexes.

Asherson points out that there may be a long delay (more than 5 years) between the first clinical features of APS and those consistent with SLE, and thus ongoing clinical vigilance for the development of new clinical features is very important in all PAPS patients. He also summarizes the various forms of secondary APS (SAPS), a term that may become redundant given the recommendations of the Sydney classification criteria and discussions from the 2007 Phospholipid Conference in Florence, Italy. The concept of seronegative APS (SNAPS) is also covered in this article, including several important factors that should be additionally considered before a patient is given this label. Finally, Asherson provides a brief summary of some of the features of catastrophic APS (CAPS), the syndrome that also bears his name. This condition is also covered in greater depth in a subsequent article of this issue of Seminars in Thrombosis and Hemostasis.

The next article by Pierangeli and colleagues summarizes the current state of knowledge regarding how aPL antibodies may predispose toward thrombosis, including the first- and second-hit model. This model proposes that the first hit involves direct priming effects of aPL antibodies on a variety of coagulation, endothelial, and complement pathways, followed by a second hit from clinical events such as infection, trauma, or surgery. The first-hit mechanisms of aPL antibodies that have been demonstrated include inhibition of activated protein C (APC); antithrombin and activated factor X (FXa); binding to platelets leading to promotion of activation and production of thromboxane B2; interaction with endothelial cells and monocytes leading to increased expression of adhesion molecules (including ICAM-1, VCAM-1, and E-selectin) and tissue factor (TF) activity; complement activation and generation of proinflammatory and prothrombotic complement split products; and binding to certain important serine proteases involved in the hemostatic and fibrinolytic pathways (e.g., plasmin, tPA). In addition, anti-β2GPI has been shown to interfere with the ability of β2GPI to inhibit von Willebrand factor (VWF)-induced platelet aggregation. Given the complexity of many of these pathways, it is helpful that the authors provided a figure illustrating how these apparently diverse effects of aPL antibodies interact.

Pierangeli and colleagues then discuss how this new knowledge might lead to novel therapies. The aim is to develop novel treatments that interfere with these first-hit mechanisms, thus reducing the risk of thrombosis when second-hit events occur. The hope is that these agents will have less toxicity than do currently available therapies (mainly anticoagulation). For example, the authors discuss how specific inhibitors of complement activation may provide a novel means to prevent vascular thrombosis in APS. In addition, a specific p38 MAPK inhibitor (SB203580) has been shown to significantly abrogate aPL–mediated platelet aggregation, and inhibition of complement activation prevented aPL–induced fetal loss and growth retardation and aPL–mediated thrombosis in a mouse model. Hydroxychloroquine and anti-GP IIb/IIIa antagonists may play a role in APS by inhibiting aPL antibodies effects on platelets. Finally, HMG-CoA reductase inhibitors (or statins) appear to reduce the procoagulant and proinflammatory effects of aPL antibodies.

A detailed discussion of how aPL antibodies may lead to infertility and fetal loss is also provided in this article. The authors mention the in vitro evidence that suggests aPL antibodies, in particular anti-β2GPI, exert a variety of effects that result in defective placentation via direct nonthrombotic effects on trophoblasts including interference with adhesion molecule expression on trophoblasts; increasing apoptosis; various effects on decidual stromal cells (including complement regulatory proteins) on the maternal side of the placenta; and perhaps a direct proinflammatory complement-mediated effect. One interesting model involves the binding of β2GPI to phospholipids on trophoblast and decidual stromal cells surfaces, which then acts as a target for aPL antibodies (in particular anti-β2GPI) and when bound mediates intracellular signaling through as yet unknown mechanisms. The binding of aPL antibodies to β2GPI on the surface of trophoblasts may also “trap” IgG isotype aPL antibodies trying to cross the placenta, thus explaining why maternal IgG aPL antibodies generally do not cause thrombotic events in fetuses or neonates. In addition to what would be considered an obvious mechanism of intraplacental thrombosis with consequent impairment of maternal-fetal blood exchange, there is now evidence that aPL antibodies may also induce a procoagulant state at the placental level. Disruption of the annexin A5 “shield” by aPL antibodies is another important proposed mechanism. Defective placentation in APS may be due to a range of direct effects of aPL antibodies on maternal decidua and invading trophoblast cells. Given that the presence of β2GPI (that bind to phospholipids) on the surface of trophoblasts is a critical component of the above models, inhibition of β2GPI binding or its removal from the trophoblast surface may be a novel method to improve pregnancy rates and outcome in APS. These models also raise the possibility that heparin may partially exert its pregnancy-prolonging effects in APS patients by binding to the phospholipid binding site on β2GPI, resulting in β2GPI being displaced from the trophoblast surface and thus no longer being available to bind to aPL antibodies.

Next, Murray Adams reviews the role that the tissue factor (TF) pathway may play in the pathogenesis of APS. This pathway is the major physiologic trigger of blood coagulation, and TF-containing microparticles may have an important role in maintaining thrombin generation during normal hemostasis. There is experimental evidence that TF activity on monocytes is stimulated by aPL antibodies from APS patients; monocytes from APS patients exhibit increased expression of TF and TF mRNA; plasma levels of TF antigen are increased in APS patients; and TF-containing microparticles are elevated in patients with positive aPL antibodies; all of which would contribute to increased TF activity and thus hypercoagulability in APS. However, there is greater uncertainty about the status of the major natural regulator of the TF pathway, Tissue factor pathway inhibitor (TPFI). There is some evidence that anti-β2GPI may bind to and/or inhibit TFPI, which would then presumably lead to greater TF activity, hypercoagulability, and an increased risk of thrombosis in APS.

In the fifth article, Asherson and colleagues provide a summary of the basic immunologic concepts related to the origin of autoantibodies in general, and that specifically relevant to aPL antibodies. These concepts include the origin of natural autoantibodies; aberrant circumstances (including oxidant stress) whereby dysregulation of the normal control mechanisms of these natural autoantibodies may occur, thereby rendering them pathogenic or pathologic; normal interactions between apoptotic cells and both β2GPI and prothrombin; and interactions between anti-β2GPI and toll-like receptors resulting in upregulation of proinflammatory cellular activity. This article then summarizes the protean clinical manifestations of APS, some of which are subsequently covered in greater detail by other authors within this issue of Seminars in Thrombosis and Hemostasis, followed by a summary of the current treatment of patients with APS.

In the next article, Tincani and colleagues review the characteristics, diagnosis, pathogenesis, and treatment of the APS in pregnancy. This is an important area of clinical medicine, given the recognized significant risk of APS to both the mother and fetus during pregnancy, especially because of the increasing number of women with APS striving to fall pregnant and deliver a live healthy child. In addition, because obstetricians now routinely include testing for aPL–related antibodies (to varying extents, but typically with at least the aCL test) in women with miscarriages, more and more women with primarily obstetric manifestations of primary APS are being identified. The authors discuss the mechanisms by which the presence of aPL antibodies not only reduces fertility but also increases the rate of early and late miscarriages. Interestingly, these mechanisms include nonthrombotic pathways, including interference with the so-called protective shield of annexin V (PAP-1) on the developing trophoblast, and complement-mediated fetal injury. Treatment of pregnant APS women still revolves around the combination of low-dose aspirin (LDA) and heparin. The authors discuss the preference of low-molecular-weight heparin (LMWH) over unfractionated heparin due to a lower incidence of thrombocytopenia and osteoporosis, especially given the otherwise prolonged exposure of the mother to heparin therapy during pregnancy. The authors also provide a timely reminder that successful pregnancy outcome in APS requires close supervision by a multidisciplinary team and that medication for APS is only one aspect of management.

In the seventh article, Beverley Hunt provides a useful summary of the features of pediatric APS. Important differences with respect to thrombotic events (in general rather than just APS associated) in children compared with adults include that such events are rare (0.07 to 5.3 per 10,000 hospital admissions), are usually provoked by the use of invasive vascular interventions in sick children, may occur in unusual sites, and are often associated with genetic prothrombotic tendencies. As a reminder of just how uncommon pediatric APS is, the Ped-APS Register has far less patients (87) than does the register for catastrophic APS (CAPS) patients discussed in a later article, which, while still considered a rare condition, contains more than 3 times (300) as many patients. Transient infection-related aPL antibodies appear to be the most common type of aPL antibodies in children, with the concomitant role of protein S deficiency being a significant cofactor when thrombosis does occur in the setting of varicella and streptococcus coinfection. The extremely wide range (2 to 82%) in reported prevalence figures of aCL in healthy children has several possible explanations, with the well-recognized variability between different aCL assays likely to be a significant factor. It is thus not surprising that due to the small number of children reported and assay (both aCL and LA testing) variability, there is conflicting data regarding the relationship between aPL antibodies and thromboembolic events in children.

Certain aspects of management also differ between children and adults and include that asymptomatic persistent (> 3 months) aPL antibodies in children generally are not treated, even with LDA, whereas LDA is used prophylactically if the child has SLE concomitant with positive aPL antibodies. These approaches are supported by the recent finding by Erkan et al that prophylactic LDA did not reduce thrombosis rates in asymptomatic adults with positive aPL antibodies, but thrombosis was more common in patients who also had SLE.[5] We are also reminded that there is a general reluctance to conduct trials with aspirin in children due to the association of this drug with Reye's syndrome. There is only limited data regarding the safety of anticoagulation in children, but the high risk of hemorrhage during play and sport has to be taken into consideration as well as compliance issues during adolescence. In children, LMWH therapy (adjusted for weight) is certainly more convenient than unfractionated intravenous heparin, both for the prophylaxis and treatment of acute thrombotic events. The optimum target international normalized ratio (INR) range for oral warfarin therapy is still debated, but a lower range of 2.0 to 3.0 is recommended by the author for initial treatment of all children with APS, as children generally have a lower prothrombotic state than do adolescents or adults. However, in adolescents with arterial thrombotic events, the author mentions that a higher INR range of 3.0 to 4.0 may be necessary, as for adults with recurrent venous thromboses and/or arterial thromboses.

Wolfgang Miesbach then covers the interesting area of malignancies and aPL antibodies. It has been recognized for almost 150 years that thrombosis occurs more frequently (up to 15% of all cancer patients) in the presence of underlying malignancy, the so-called Trousseau's syndrome. In addition to procoagulant effects directly due to the malignancy, the use of chemotherapy and the need for indwelling vascular access further increases the risk of thrombosis in these patients. Interestingly, thrombosis may also be the first manifestation of an underlying malignancy with an odds ratio of 4.8 for the development of cancer in patients with idiopathic versus secondary thrombosis. A wide range of solid and hematologic cancers has been described with positive aPL antibodies (mainly in case reports). Epidemiologic studies have described a higher frequency of positive aCL in cancer patients versus healthy controls (22% vs. 3% in one study), with the aCL-positive patients also having a significantly higher incidence of thrombosis. It appears that solid malignancies may have a stronger association with positive aPL antibodies and clinical manifestations of APS than do hematologic malignancies, but this requires further study.

Conversely, patients with APS may also be at increased risk of developing cancer, in particular carcinomas. The presence of an underlying malignancy may also be a risk factor for CAPS, the most severe form of APS. Nine percent of CAPS patients present with an underlying cancer, and these patients were generally older and had the worse prognosis of the whole CAPS cohort. An interesting observation that LMWH therapy may have antineoplastic effects raises the question of whether aPL antibody–positive patients with underlying malignancy should be treated prophylactically with LMWH rather than with oral anticoagulants. However, prolonged use of LMWH is less convenient than use of oral warfarin and carries an increased risk of heparin-induced osteoporosis and heparin-induced thrombocytopenia and thrombosis syndrome (HITTS), albeit less so than with unfractionated heparin. In medicine, exceptions always occur and the author also reports that some cases of non–Hodgkin's lymphoma (NHL)-related aCL are not associated with an increased risk of thrombosis, but this may reflect the known weaker association between IgM aCL (even in high titer) and thrombotic manifestations in APS. However, it remains uncertain whether aPL antibodies develop as an epiphenomenon of malignancy and the extent to which positive aPL antibodies predispose to thrombosis in patients with cancer.

In the second article from this author, Wolfgang Miesbach provides a concise review of neurologic manifestations of APS, where the underlying pathology is typically a nonvasculitic thrombotic occlusion of cerebral vessels but nonischemic/nonthrombotic manifestations are also described. He suggests that the high frequency of neurologic manifestations in APS may reflect an increased susceptibility of the cerebral vasculature to APS, possibly as a result of the higher phospholipid content of neuronal tissue and endothelium. It is not surprising that aPL antibodies have been suggested as an important risk factor for stroke, especially in younger (under the age of 50 years) patients, with a reported frequency ranging between 18% and 43%. Data from the Framingham cohort suggests that positive aCL results are an independent risk factor for future ischemic stroke and transient ischemic attacks (TIA) in women but not men.[6] Looking at this issue from another perspective, strokes are the most common neurologic manifestation in patients with APS, occurring in 42 of the 100 APS patients reported by Munoz-Rodriguez and colleagues.[7]

With respect to nonischemic/nonthrombotic neurologic manifestations, the author reports that such manifestations occur in 40% of unselected patients with acute neurologic symptoms and positive aPL antibodies. These include several nonischemic neurologic disorders in which the evidence for a definite association with APS is still inconclusive and warrants further study. Such disorders include migraines, multiple sclerosis, epilepsy (in the absence of cerebral infarction), and dementia. Mechanisms whereby aPL antibodies directly bind to specific neuronal cells (including synaptoneurosomes) have been proposed for neurologic conditions such as chorea and transverse myelitis. Others such as cognitive dysfunction (a common feature in APS) are probably multifactorial and include ischemic etiologies.

The severe and often life-threatening form of APS is covered extensively by Cervera and colleagues in the next article. Described initially as the catastrophic antiphospholipid antibody syndrome (CAPS), the term Asherson's syndrome has also been used to honor Asherson's initial report of this syndrome in 1992,[8] as well as his ongoing involvement in further defining this syndrome, and largely in collaboration with Cervera's group in Barcelona. Cervera's group now maintain the Internet-accessible CAPS Registry. Freely available, the primary utility of the registry is an ability to pool the clinical experience of an otherwise rare condition that any individual clinician or unit might only see once or twice per year. Accordingly, this registry has enabled the publication of various insights into the frequency of differing clinical manifestations as well as the various treatment modalities used in these frequently critically ill patients. At the present time, the combined use of anticoagulation (AC), corticosteroids (CS), and plasma exchange (PE) is recommended as first-line therapy. Many clinicians substitute intravenous immunoglobulin (IVIG) for plasma exchange due to lack of facilities to perform plasma exchange. The registry suggests a slightly lower recovery rate of 69% for AC + CS + PE and/or IVIG compared with 78% for AC + CS + PE, but given the relatively small patient numbers, the confidence intervals for these values are likely to show some overlap. The data also reminds us of the improving, but still high, mortality rate in patients with CAPS, with cerebral involvement remaining the main cause of death. In the period between 2001 and February 2005, approximately one third of patients with CAPS still died, emphasizing the limitations of current therapeutic measures and the need for both earlier recognition of the syndrome and more potent treatments.

The final article in this issue is by the guest editors. We discuss several important aspects of the diagnosis of APS, which may present a challenge for a general practitioner or any other clinician who is not intimately familiar with APS, particularly if the presentation does not include one of the “classic” presentations of significant venous and/or arterial thrombosis and/or pregnancy morbidity. This includes patients with seronegative APS (SNAPS), and we discuss how the diagnosis of this subgroup of APS may be affected by variations in testing for aPL antibodies (aCL, anti-β2GPI, LA, and other tests) and recent proposals regarding which aPL tests should be routinely performed. We also revisit the role of the classification criteria for APS, in particular that they were designed primarily to have high specificity (but at the potential cost of sensitivity) thus enabling the correct identification of APS patients for inclusion into prospective studies rather than as a set of clinical diagnostic criteria. The importance of this distinction is that should clinicians apply the criteria strictly when trying to make a diagnosis of APS, some patients with APS may be missed and thus inappropriately managed.

We thank all the contributors to this issue of Seminars in Thrombosis and Hemostasis for their excellent and timely contributions. We hope that you, as reader, find the contents both informative and enjoyable. Look out for the next issue of Seminars in Thrombosis and Hemostasis, where the APS story continues.

Note added in proof: The guest-editors of this issue of Seminars in Thrombosis and Hemostasis were saddened to hear of the recent passing of Professor Ronald A. Asherson in May 2008. Professor Asherson was a true pioneer in the field of antiphospholipid syndrome and continued his passion for the study of this disease until his sudden death. We wish to dedicate this issue of Seminars in Thrombosis and Hemostasis to his memory and in tribute to his contribution to this field.

REFERENCES

• 1 Harris E N, Gharavi A E, Boey M L et al.. Anticardiolipin antibodies: detection by radioimmunoassay and association with thrombosis in systemic lupus erythematosus.  Lancet. 1983;  2 1211-1214
• 2 Wilson W A. Classification criteria for antiphospholipid syndrome.  Rheum Dis Clin North Am. 2001;  27 499-505
• 3 Miyakis S, Lockshin M D, Atsumi T et al.. International consensus statement on an update of the classification criteria for definite antiphospholipid syndrome (APS).  J Thromb Haemost. 2006;  4 295-306
• 4 Favaloro E J, Wong R C. Antiphospholipid antibodies and the antiphospholipid syndrome II: limitations, standardization, and clinical utility of laboratory testing.  Semin Thromb Hemost. 2008;  34 , (in press)
• 5 Erkan D, Harrison M J, Levy R et al.. Aspirin for primary thrombosis prevention in the antiphospholipid syndrome: a randomized, double-blind, placebo-controlled trial in asymptomatic antiphospholipid antibody-positive individuals.  Arthritis Rheum. 2007;  56 2382-2391
• 6 Janardhan V, Wolf P A, Kase C S et al.. Anticardiolipin antibodies and risk of ischemic stroke and transient ischemic attack: the Framingham cohort and offspring study.  Stroke. 2004;  35 736-741
• 7 Muñoz-Rodriguez F J, Font J, Cervera R et al.. Clinical study and follow-up of 100 patients with the antiphospholipid syndrome.  Semin Arthritis Rheum. 1999;  29 182-190
• 8 Asherson R A. The catastrophic antiphospholipid syndrome.  J Rheumatol. 1992;  19 508-512