Update of the guidelines for lupus anticoagulant detection

Abstract

Summary.  One of the conclusions of the subcommittee meeting on Lupus Anticoagulant/Phospholipid dependent antibodies, held in Geneva on 2007, was the need to update the guidelines on Lupus Anticoagulant (LA) detection. Particular emphasis was given to several aspects discussed in this official communication. A new paragraph is dedicated to the patient selection, and aims to minimize inappropriate requests for LA testing. Modalities for blood collection and processing are fully delineated and the choice of tests is limited to dRVVT and a sensitive aPTT. Calculation of cut-off values for each diagnostic step are clearly stated. A final paragraph reports the interpretation of the results in general and in particular situations.

These guidelines are intended to update the criteria for the detection of the presence of lupus anticoagulants (LA) that were originally proposed by Brandt et al. in 1995 [1]. The subcommittee on Lupus Anticoagulant/Phospholipid-dependent Antibodies acknowledges that the present guidelines have been extremely useful during the past 13 years but that it is now appropriate to provide additional details and specifications in light of the knowledge and experience that has been accumulated since their publication.

Patient selection

Testing for LA should be limited to patients who have a significant probability of having the antiphospholipid syndrome (APS), or who have unexplained prolonged aPTT in the course of routine laboratory testing. Appropriateness to search for LA can be graded according to clinical characteristics into low, moderate and high. Low: venous (VTE) or arterial thromboembolism in elderly patients; Moderate: accidentally found prolonged aPTT in asymptomatic subjects, recurrent spontaneous early pregnancy loss, provoked VTE in young patients; and High: unprovoked VTE and (unexplained) arterial thrombosis in young patients (< 50 years of age), thrombosis at unusual sites, late pregnancy loss, any thrombosis or pregnancy morbidity in patients with autoimmune diseases (systemic lupus erythematosus, rheumatoid arthritis, autoimmune thrombocytopenia, autoimmune hemolytic anaemia) [2]. Generalized searches performed on blood samples obtained from asymptomatic individuals or categories of patients other than described here are highly discouraged to avoid the risk of obtaining false-positive results that are relatively common on account of the poor specificity of the assays. Once a patient has been identified as positive for LA, it is imperative that testing be repeated on a second occasion > 12 weeks after the initial testing. It is preferable for samples to be obtained prior to, or in the absence of, anticoagulant therapy as this may interfere with the assay. Recommendations of the subcommittee are summarized in Table 1. How to determine and interpret cut-off values is described in Table 2.

Table 1.   Recommendations for the optimal laboratory detection of lupus anticoagulant (LA)*

*Further explanations and qualifications are reported in the text. PNP, pooled normal plasma; TT, thrombin time; PL, phospholipids

(A) Blood collection

1. Blood collection before the start of any anticoagulant drug or a sufficient period after its discontinuation

2. Fresh venous blood in 0.109 m sodium citrate 9:1

3. Double centrifugation

4. Quickly frozen plasma is required if LA detection is postponed

5. Frozen plasma must be thawed at 37 °C

(B) Choice of the test

1. Two tests based on different principles

2. dRVVT should be the first test considered

3. The second test should be a sensitive aPTT (low phospholipids and silica as activator)

4. LA should be considered as positive if one of the two tests gives a positive result

5. For interpretation see Table 2 (screening test)

(C) Mixing test

1. PNP for mixing studies should ideally be prepared in house. Adequate commercial lyophilized or frozen PNP can alternatively be used

2. A 1:1 proportion of patient : PNP shall be used, without preincubation within 30 min.

3. LA can not be conclusively determined if the TT of the test plasma is significantly prolonged

4. For interpretation see Table 2 (mixing test)

(D) Confirmatory test

1. Confirmatory test(s) must be performed by increasing the concentration of PL of the screening test(s)

2. Bilayer or hexagonal (II) phase PL should be used to increase the concentration of PL.

3. For interpretation see Table 2 (confirmatory test)

(E) Expression of results

1. Results should be expressed as ratio of patient-to-PNP for all procedures (screening, mixing and confirm)

(F) Transmission of results

1. A report with an explanation of the results should be given

Table 2.   Cut-off values for lupus anticoagulant (LA) detection*

Cut-off values
*Testing described above must be performed with the local reagent/coagulometer combination on plasmas from at least 40 adult healthy donors less than 50 years of age. Do not use cut-off values established elsewhere even although they refer to the same method and coagulometer. ^†The clotting time of the confirmatory test in LA positive samples is not always shortened to within the normal range of controls. To avoid false-negative results, the Subcommittee recommends confirmatory tests to be performed in all the normal controls and to use the mean of obtained clotting times to calculate the percentage of shortening. This percentage can be used as a cut-off value. ICA, index of circulating anticoagulant.
Screening test
How should this be determined
1. Perform testing on plasmas from healthy donors
2. Take the cut-off as the value above the 99th percentile of the distribution
Interpretation
1. Results of screening tests are potentially suggestive of LA when their clotting times are longer than the local cut-off value
Mixing test
How should this be determined
1. Perform testing on plasmas from healthy donors mixed with the pooled normal plasma (PNP) at 1:1 proportion. Testing should be performed without pre-incubation within 30 min
2. Take the cut-off as the value above the 99th percentile of the distribution
3. Alternatively, the cut-off may be the value of the ICA defined according to the equation: ICA = [(b − c)/a] × 100, where a, b and c are the clotting times of the patient plasma, mixture and normal plasma, respectively [16]
Interpretation
1. Results of mixing tests are suggestive of LA when their clotting times are longer than the local cut-off value, or when the ICA is greater than the local cut-off value
Confirmatory test
How should this be determined
1. Perform testing on plasmas from healthy donors at low (screen) and high (confirm) phospholipid concentration
2. Take the cut-off as the value corresponding to the mean of the individual % corrections calculated as defined by the equation [(screen − confirm)/screen] × 100^†
Interpretation
1. Results are confirmatory of LA if the % correction is above the local cut-off value

Explanation and clarification of the recommendations summarized in Table 1

Numerous variables can affect assays used for LA detection. Among them, the content and type of phospholipids (PL) in the reagent mixture, activator, plasma preparation, expression of results and cut-off values greatly influence the results. Moreover, as the spectrum of antibodies and their epitope specificity may vary widely, reference material constitutes a major problem. As a consequence a high variability in the performance of clinical laboratories with respect to sensitivity and specificity of LA tests has recently been reported [3–7]. The rates of false-positive and false-negative detections remain relatively high. The former are of particular concern because they qualify the patients for long and unnecessary oral anticoagulant treatment [7].

(A) Blood collection

A3: Double centrifugation of the sample should be performed to ensure that the sample is platelet poor. This can be achieved by transferring the plasma after the initial centrifugation process (2000 g, 15 min, room temperature) to a non-activating plastic centrifuge tube using a plastic pipette, then recentrifuging the plasma for an additional 10 min at a higher speed (> 2500 g). When aliquoting to a secondary tube, take care to not include the residual platelets that may have collected at the bottom of the centrifuge tube [8]. Plasma filtration is not recommended as this introduces variables (type of filter, amount of plasma to be filtered, loss of von Willebrand factor and added costs [9]. Moreover, the problems of filters availability and adjunctive costs must be considered.

A4: Freezing of the samples must be performed as quickly as possible after venipuncture to prevent loss of coagulation factors. Freezing the plasma in a freezer at −70 °C or below is reasonable.

A5: Before analysis, frozen plasma must be thawed by total immersion of sample content in a water bath at 37 °C for 5 min to avoid formation of cryoprecipitate and then mixed thoroughly before testing.

(B) Choice of the test

B1: The risk of false-positive results is increased to an unacceptable level if more than two screening tests are performed.

B2: There is evidence that no single test is sensitive for all LA. The recommendation is to perform two different tests that represent different assay principles. Diluted Russell Viper Venom time (dRVVT) is widely used in clinical laboratories and is believed to be specific for detecting LA in those patients at high risk of thrombosis [10]. An international External Quality Assessment Programme for laboratories working in the field of thrombosis showed that dRVVT is the most robust test in detecting LA [11].

B3: Any aPTT test performed with silica as an activator and low PL content is the second test of choice because of its sensitivity [3,6] for LA. Kaolin as an activator is not recommended because of its problematic behaviour in automated coagulometers. Ellagic acid as an activator is not recommended as a result of its insensitivity for LA. The Subcommittee does not recommend the following tests: dilute prothrombin time (dPT) because of variability in thromboplastin reagents; assays based on snake venoms such as Ecarin and Textarin because there are no standardized commercial assays available based on this principle; Kaolin Clotting Time (KCT) as a result of its poorer reproducibility compared with the other tests available [11].

(C) Mixing test

C1: The pooled normal plasma (PNP) should be prepared ‘ad hoc’ (home-made) by double centrifugation to ensure that the PNP contains minimal residual platelets (< 10⁷ mL⁻¹) and to ensure approximately 100% activity for all clotting factors. The material must be stored frozen (−70 °C) in small aliquots. Commercial lyophilized or frozen normal plasmas can be used if they fulfill the above specifications or have otherwise been validated for use in LA testing. When testing for LA during pregnancy, normal range(s) of clotting times defined for normal pregnant women should be considered (aPTT is, in general, shortened as a result of high factor VIII levels during pregnancy, but also dRVVT can change for unknown reasons). No reference plasma is available. Therefore, established positive and negative plasmas should be used as controls to validate the assay.

C3: The coagulation time of a mixing test could also be prolonged because of the presence of heparin or inhibitors to coagulation factors. The thrombin time (TT) performed on test plasma or the clinical history of bleeding will help to identify heparin or specific inhibitors to clotting factors, respectively. LA screening is not possible if the test plasma is unclottable or if the content of heparin in the test plasma exceeds the reagent neutralization capacity. There are commercial dRVVTs and APTTs containing neutralizers able to quench heparin up to 0.8 U mL⁻¹. Although limited experience is available on the effect of low-molecular-weight heparin (LMWH), screening for LA in patients treated with LMWH is possible. It should, however, be noted that the effect on LA assays may vary depending on the ratio between factor (F) Xa to FIIa activity of each LMWH preparation. The effect of direct thrombin or FXa inhibitors on LA assays is unknown. Whereas hydroxychloroquine may weakly interfere with LA testing directly affecting the formation of IgG-ß2GPI complexes on phospholipid bilayers [12], aspirin and clopidogrel do not interfere.

(D) Confirmatory test

D2: Freeze/thawed platelets are not recommended as the source of PL for the confirmatory tests because of poor batch-to-batch consistency.

(F) Transmission of results

F1: LA test results should be reported with quantitative results, and an interpretative comment that indicates whether the findings are compatible with the presence/absence of LA. This is important as many clinicians may not be aware of the significance of all the complex testing procedures carried out by the laboratory. Comments such as borderline or dubious LA are highly discouraged and in these cases the comment should be limited to the following: ‘to be tested again in 1 week’.

Interpretation of results

Integrated tests

Integrated tests include screening and confirmation in a single procedure. Such tests consist of testing the plasmas under investigation twice by means of the dRVVT [13] or APTT [13,14] performed in parallel at low (screen) and high (confirm) phospholipid concentrations. In principle, these tests do not require performance of the mixing test and the results may be interpreted according to the specific cut-off values by calculating either the percentage correction [(screen − confirm)/screen × 100] [13] or the LA ratio (screen/confirm) [15]. Both the percentage correction and the LA ratio may benefit from normalization of results against a PNP run in parallel with the test plasmas [(screen/confirm) patient divided by (screen/confirm) PNP]. Some of the above tests are designed to measure the coagulation times on a mixture of patient and PNP [14].

LA detection during acute thromboembolic events

Caution should be exercised in interpretation of the results of tests performed close to a thromboembolic event as patients may be treated with full doses of unfractionated heparin and/or vitamin K antagonists (VKA). Furthermore, acute-phase reactants as FVIII may be increased during acute events.

Antiphospholipid antibody profiles

A LA result should always be considered in the context of a full laboratory aPL profile comprising anticardiolipin (aCL) and antiβ2glycoprotein I (aβ2GPI) antibodies ELISAs. The presence of medium-high titers aCL and aβ2GPI of the same isotype (most often IgG) is in agreement with a positive LA and identifies patients at high risk for thrombosis. Less information is available for the correlations with fetal losses. Isolated LA positivity is significantly more frequent in subjects without clinical events [19] or may be false-positive especially if identified as ‘mild in potency’, if it is found in elderly patients or if it is diagnosed for the first time [7].

LA detection in patients on long-term vitamin K antagonists (VKA)

• 1
   The interpretation of results is difficult because of the prolonged basal clotting time. To avoid misinterpretation, it is recommended to perform laboratory procedures 1 to 2 weeks after discontinuation of treatment or when the international normalized ratio (INR) is less than 1.5. Bridging VKA discontinuation with LMWH is recommended with the last dose of LMWH administered more than 12 h before the blood is drawn for LA testing.
• 2
   Alternatively, if the INR is between 1.5 and < 3.0, a 1:1 dilution of patient plasma and PNP can be considered. Note, that the interpretation of results may still be difficult and that the LA titer will be diluted 2-fold.
• 3
   Detection procedures by Textarin(Taipan)/Ecarin clotting times [17,18] or integrated tests (i.e. % correction for APTT, SCT and dRVVT at low and high phospholipid concentration) [13] are not currently recommended as they require further critical evaluation.

Acknowledgements

The authors would like to thank the following colleagues who helped in writing these guidelines with suggestions and comments: Katrien M.J Devreese, Ghent, Belgium; Annie Robert, Paris, France; Alicia N Blanco, Buenos Aires, Argentina; Beatriz E Grand, CABA, Argentina; Ricardo Forastiero, Buenos Aires, Argentina; Tatsuya Atsumi, Sapporo, Japan; Galit Sarig, Haifa, Israel; E Favaloro, Westmead, Australia; Veena Chantarangkul, Milano, Italy.

Disclosure of Conflict of Interests

The authors state that they have no conflict of interest.