Antiphospholipid Antibodies and the Antiphospholipid Syndrome II: Limitations, Standardization, and Clinical Utility of Laboratory Testing

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

The opening article is by Pierangeli and Harris, who ask the pertinent question: A quarter of a century in anticardiolipin antibody testing and attempted standardization has led us to here, which is? The anticardiolipin (aCL) test has been widely used by physicians for diagnosing patients with APS since the mid-1980s. An appropriate diagnosis enables the effective management of patients with recurrent thrombosis and recurrent pregnancy losses. The test was first established in 1983 as a radioimmunoassay[1] and soon thereafter converted into an enzyme-linked immunosorbent assay (ELISA). The other test commonly used in the diagnosis of APS is the lupus anticoagulant (LA) test. The aCL ELISA is sensitive for the diagnosis of APS but unfortunately lacks specificity. On the other hand, the LA assay, although more specific, is not as sensitive as the aCL ELISA. Other specific tests for APS are now also available such as the anti-β2 glycoprotein I (anti-β2GPI) assay, the antiprothrombin (anti-PT) assay, and other ELISAs that use negatively charged phospholipids instead of cardiolipin to coat the plates. Again, these tests, though specific for APS, are not as sensitive for the disorder as the aCL.

In the past 25 years, there have been numerous efforts to standardize aCL, LA, and anti-β2GPI tests, but there are still significant issues related to both intralaboratory and interlaboratory variation in results for all three assays. Pierangeli and Harris discuss in detail the clinical value of these tests, technical problems associated with their use, the current laboratory classification criteria for APS, and possible new and better assays that will be available in the near future to assist in the laboratory diagnosis of APS.

In the second article of this issue of Seminars in Thrombosis and Hemostasis, Monica Galli continues the discussion regarding the clinical utility of laboratory tests used to identify aPL and diagnose APS. The author reminds us that APS is defined by thrombotic and/or obstetric events together with the presence of aPL in plasma of patients. The original laboratory criteria of APS in the 1999 Sapporo criteria[2] included LA and/or IgG/IgM aCL. They were recently updated in the 2006 Sydney criteria[3] with the addition of IgG/IgM anti-β2GPI, a better definition of “medium to high antibody titer,” and the extension to 12 weeks in terms of fulfilling the requirement regarding the “persistence in time,” but removal of the requirement that the aCL assay has to be performed in a β2GPI-dependent manner.[3] The author asserts that whereas the revised criteria represent an overall improvement, they increase the potential for the “overdiagnosis” of APS, thus putting patients at risk of inappropriate treatment with anticoagulation. To reduce this possibility, proposals have been put forward to implement strict guidelines for the performance of the LA assay, to exclude aCL measurements in their current application from the criteria, and to limit the measurement of anti-β2GPI to the IgG isotype.[4] The author contends that these proposals should also help to simplify the laboratory workup of patients being investigated for APS. Nevertheless, some of these proposals are contentious, and other authors within this issue of Seminars in Thrombosis and Hemostasis believe that too liberal an adoption of such proposals could lead to the alternative adverse situation of “underdiagnosis” of APS, and thus undertreatment of some individuals with clinically defined APS.

The next article discusses what may potentially become the fourth aPL assay. Oku and colleagues evaluate the detection and clinical utility of antiprothrombin antibody testing. They begin by reminding us that aCL, anti-β2GPI, and LA are the only laboratory tests considered within the revised criteria for the classification of APS. However, antibodies against phosphatidylserine-prothrombin complex (aPS/PT) have recently been detected, and these antibodies, rather than antibodies against prothrombin alone, are more closely associated with APS and LA. The sensitivity and specificity of aPS/PT for the diagnosis of APS were assessed by the authors in a population of patients with a variety of autoimmune disorders. They determined that aCL antibodies and aPS/PT have similar diagnostic value for APS, and thus propose that aPS/PT assay should be further explored, not only for research purposes but also as a candidate for one of the laboratory criteria for the classification of the APS.

Reber and colleagues then ask another pertinent question in this interesting field, namely: Is the standardization of technical aspects in laboratory testing for aPL an impossible dream? They begin their story by reminding us that aCL and anti-β2GPI assays are widely performed because they are part of the laboratory criteria for the identification of APS. Despite several standardization workshops and the availability of a worldwide accepted calibrator material for aCL, a high variability in numerical assay results and sample classification is still reported in comparative studies and external quality control surveys. For anti-β2GPI assays, comparison of numerical values is impeded by the absence of a common calibrator material, and external quality surveys similarly show a large overlap in sample classification. Numerous variables impact assay results, among them the source and integrity of β2GPI, the secondary calibration process, and the assessment and derivation of cutoff values. The vast majority of laboratories use commercial kits for both assays, and the number of available kits has risen considerably in past years. Nevertheless, many problems persist, and there is a need to improve the comparability in assay results. Reber and colleagues believe that the use of monoclonal antibodies as reference calibrators has to be especially considered, with their suitability evaluated by future collaborative studies and external quality control surveys. This proposal is tempered by the comments of other contributors to this issue of Seminars in Thrombosis and Hemostasis who point out several limitations to such an approach. Even so, it is correct to raise the possibility and then to evaluate the proposal in appropriate studies. The authors of this article also assert that kit manufacturers should provide more precise information on results obtained when testing control groups for establishing reference ranges and cutoff values and that each laboratory, even if using commercial kits, assess its local cutoff value whenever possible. They conclude by surmising that in the field of autoimmunity, assay standardization is a difficult but nevertheless important task and that much more effort is needed to reduce the high interlaboratory variability in assay results even if absolute standardization cannot be feasibly achieved.

In the next article, de Groot and co-workers continue the story regarding ongoing failures to achieve standardization of aPL assays. They assert that confusion concerning who has and has not APS continues to plague us in the year 2008, some 22 years after Graham Hughes first introduced the term antiphospholipid syndrome [5] to characterize a group of patients with thrombosis and/or recurrent complications of pregnancies with the presence of aPL within their plasma. Although to their mind the clinical criteria are well defined, there remains an ongoing issue regarding serologic criteria, and lack of standardization of the assays that define the serologic criteria, notably phospholipid-dependent coagulation assays such as LA, and ELISA for aCL and anti-β2GPI. This has led to heated arguments regarding which population(s) of antibodies should be measured to detect a patient at risk for (recurrent) thrombosis or pregnancy complications.

de Groot and colleagues believe that everybody agrees on the need to better standardize the assays, but different views are held on how this should be achieved and that commercial considerations have hampered consensus on which assays should be applied, how they should be performed, and the cutoff values that discriminate between pathologic and nonpathologic results. New prospective cohort studies to reevaluate the clinical significance of the available assays are essential, but the relatively limited numbers of APS patients visiting single hospital departments or units means that multicenter studies will be required, with their inherent complexity in organization. The article discusses the shortcomings of the current serologic assays, provides some potential strategies to solve these shortcomings, and discusses new developments/assays to improve the specificity of such assays for thrombosis and pregnancy complications.

In the next article, Andreoli and colleagues discuss the concept that emerging technologies in aCL detection might signal a shift in the direction or focus of current attempts at standardization of aPL testing. Currently, aPL are mostly assayed using commercial ELISA kits, whose performance has not been found to be sufficiently concordant among the different manufacturers. In the past years, collaborative groups have spent considerable effort to reach some form of standardization, but this process is still ongoing. Such lack of standardization has recently become even more crucial, as manufacturers have had to face an increasing demand for fully automated tests for aPL, like those test systems that have been developed for other autoantibodies (e.g., anti-dsDNA antibodies, MPO-ANCA, and PR3-ANCA). With the advent of new technology, should current attempts at standardization be abandoned and refocused to new technologies? Or will the new technologies just raise additional problems to the process of standardization? The authors thus report their recent experience with two newly developed automated methods for aCL testing. In particular, the evaluations were performed using routine samples, as they believe that these better reflect the “real-life” situation in which these automated methods will operate. They also mention other “emerging technologies” in the field of aPL detection.

The next article by the guest editors provides a summary of the consensus process as applied to the development of guidelines for the laboratory testing and reporting for aPL assays, in particular the recent initiatives in Australia and New Zealand via the Australasian Anticardiolipin Working Party (AAWP). Wong and Favaloro discuss the pros and cons of the consensus process related to the formulation of guidelines. They point out in the area of aPL testing that due to the paucity of good-quality published evidence, there is a heavy reliance on expert opinion, and thus the guidelines for aPL testing and reporting are largely “eminence based” rather than “evidence based.” With such an approach comes some potential for biased recommendations that reflect the personal preferences of those who have the “loudest voice” or most “strongly expressed” opinions. The authors therefore share the dilemmas and decisions faced by the AAWP during the writing of their consensus guidelines for aCL[6] and anti-β2GPI[7] testing and reporting, as well as the strategies employed to minimize potential bias during the guideline formulation process. They communicate that this process is time consuming and may sometimes produce non-ideal or “watered down” recommendations but nevertheless remains an important initiative given the paucity of good-quality published evidence in this area.

The next two chapters provide a change of focus, as discussion turns to the role of LA testing within the identification of APS. LA currently appears to be the most specific assay for APS and has the strongest association with thrombosis risk.[8] Tripodi begins by discussing the diagnostic criteria and use of screening, mixing, and confirmatory studies for laboratory testing for LA. The presence of LA in plasma is a strong risk factor for thromboembolism and fetal loss associated with APS. Because of this, demand for testing for LA is increasing, but the performance of clinical laboratories in terms of its detection is still a matter of concern according to the author. This is due in part, he explains, to the lack of specific tests, but also to the lack of standardization and application of the diagnostic criteria. Accordingly, Tripodi reviews the diagnostic criteria and the many issues that may affect test results.

Ledford-Kraemer extends these concepts further in her article on the preexamination variables, mixing studies, and diagnostic criteria for LA. The author begins by reminding us that the latest international guidelines for the laboratory diagnosis of LA date back to 1995. She asserts that they are in dire need of revision; a sentiment that the guest editors and all contributors to this issue of Seminars in Thrombosis and Hemostasis most likely share. In the intervening period, various patterns of practice and differential interpretations by clinical laboratories have altered the intent or usage of the criteria and recommendations made at that time. Moreover, Ledford-Kraemer contends that various topics in the preexamination phase of LA testing (such as the role of the clinician in test ordering, the impact of direct thrombin inhibitors or oral anticoagulant therapy, and patient biologic variation) need to be considered in future guidelines and that examination issues relating to mixing studies will also need critical review. Finally, mixing studies have been applied to a variety of low-phospholipid LA screening assays, and no uniformity or standardization exists as to how these test results should be interpreted or if in fact they are suitable for identifying weak lupus anticoagulants.

The final article in this issue of Seminars in Thrombosis and Hemostasis, and indeed the final word in the story of APS for these two special issues covering APS, is from the guest editors. We discuss the fact that aPL are identified using a large number of laboratory procedures based on one of two distinct test processes, namely solid-phase assays and liquid-phase assays. The former includes aCL and anti-β2GPI antibodies, and the latter are centered on clot-based tests that are used to identify the so-called lupus anticoagulant. Our article provides an overview of the laboratory testing and identification of aPL, and in particular the limitations, standardization, and clinical utility of such testing. We also review preanalytical, analytical, and postanalytical issues that compromise the clinical utility of these tests and provide a list of recommendations aimed to foster broader international cooperation to assist in the preparation of integrated guidelines for both solid-phase and liquid-phase tests and for both laboratory testing and for clinical ordering and interpretation.

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 content both informative and enjoyable. Indeed, we also once thank all the contributors to the previous issue of Seminars in Thrombosis and Hemostasis. The guest editors have very much enjoyed partaking in this journey down the path of one of the most interesting clinical disorders known. Graham Hughes, one of the key figures in the discovery of APS, recently quoted the dean of medicine at Barcelona University as saying that APS was one of the two “new” diseases of the late 20th century (the other being AIDS).[9] As such, APS, and the role of the laboratory in its diagnosis will continue to both interest and frustrate us for many more years to come.

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 Galli M, Reber G, de Moerloose P, de Groot P G. Invitation to a debate on the serological criteria that define the antiphospholipid syndrome.  J Thromb Haemost. 2008;  6 399-401
• 5 Hughes G R. The anticardiolipin syndrome.  Clin Exp Rheumatol. 1985;  3 285-286
• 6 Wong R CW, Gillis D, Adelstein S et al.. Consensus guidelines on anti-cardiolipin antibody testing and reporting.  Pathology. 2004;  36 63-68
• 7 Wong R C, Favaloro E J, Adelstein S et al.. Consensus guidelines on anti-beta 2 glycoprotein I testing and reporting.  Pathology. 2008;  40 58-63
• 8 Galli M, Luciani D, Bertolini G, Barbui T. Lupus anticoagulants are stronger risk factors of thrombosis than anticardiolipin antibodies in the antiphospholipid syndrome: a systematic review of the literature.  Blood. 2003;  101 1827-1832
• 9 Hughes G R. Hughes syndrome. The antiphospholipid syndrome—a clinical overview.  Clin Rev Allergy Immunol. 2007;  32 3-11