Diagnosis of antiphospholipid syndrome may have far-reaching consequences with respect to patient management, and accurate laboratory screening for lupus anticoagulant (LA) is therefore crucial. Guidelines have been developed by expert bodies for LA testing in an attempt to standardize and improve the accuracy of these tests [1–3]. Several groups have shown testing for LA to be suboptimal [4,5], and we have previously shown that compliance with guidelines improves the diagnostic accuracy of laboratory tests for LA . The most recent ISTH/SSC guidelines on LA testing were published in October 2009 . A UK NEQAS exercise for LA, in April 2010, sought to investigate whether laboratories were employing methodology compliant with the recommendations given in these guidelines. A proficiency testing exercise and a methodology questionnaire were distributed to 300 laboratories, and responses were received from 295 centers in 21 countries. The proficiency testing component of the exercise comprised two samples, one from a subject with mild factor XI deficiency (FXI coagulant activity of 20 IU dL−1), considered to be LA-negative, and the other from a subject known to test strongly positive for an LA.
Several of the recommendations in the guidelines refer to preanalytic procedures, which are not assessed in proficiency testing exercises. Laboratory practice could, however, be evaluated through a questionnaire. Table 1 summarizes recommendations from the guidelines that were evaluated in this exercise, and indicates the proportion of centers that were non-compliant.
|Guideline recommendations||Proportion of centers not compliant (%)|
|Testing should be limited to patients with significant probability of having antiphospholipid syndrome, or an unexpected prolonged APTT||8*|
|Blood to be collected into 109 m of sodium citrate||9|
|Blood should be double centrifuged, not filtered||16|
|DRVVT should be the first test considered. APTT should be the second test. Further tests should not be employed||52|
|The APTT activator should be silica-based||20|
|Cut-offs for screening and mixing tests should be the 99th percentile of results from normal clotting times||89|
|Antiphospholipid antibody (APA) profiles should be considered||54†|
|Reports should include an interpretative comment indicating whether results are consistent with the presence or absence of LA||7|
The guidelines usefully introduce the concept of pretest evaluation of the likelihood of LA, directing the priority given to carrying out the investigation. Moderate likelihood is attributed to, among other features, coincidental prolonged activated partial thromboplastin time (APTT) in asymptomatic patients. However, UK NEQAS APTT data indicate that 240 of 826 (29%) of laboratories in the UK NEQAS coagulation screening test program employed an LA-insensitive reagent as their APTT reagent for routine investigations. Consequently, these laboratories risked failing to obtain prolonged APTTs resulting from LA. Whether the risk of thrombotic events in this group of patients warrants additional testing with a sensitive APTT reagent is debatable, but these data do indicate a cohort of laboratories unable to identify a subgroup of patients considered to have a moderate likelihood of LA.
A total of 295 centers responded to the survey questionnaire. Some participants did not complete all questions, so the total number of responses for each question is indicated. With respect to preanalytic methodology, the following departures from the guidelines were noted: 27 of 295 (9%) used 0.129 m citrate anticoagulant against a recommendation to use 0.109 m; 17 of 295 (6%) did not carry out double centrifugation (although two reported checking of platelet counts); and a further 30 of 295 (10%) filtered plasma prior to storage, which the ISTH guideline states is not recommended .
The ISTH/SSC guidelines recommend that the dilute Russell viper venom time (DRVVT) and APTT should be the main tests considered for LA testing. Furthermore, the increased likelihood of false-positive LA test results, with consequences for patient treatment, has led to the recommendation not to perform more than these two tests. The UK NEQAS exercise revealed that 136 of 286 (48%) centers used APTT and DRVVT only, with a similar proportion of centers performing one or more additional tests. Over 50% of centers did not comply with the recommendations to employ only APTT and DRVVT. The guidelines do not recommend the use of dilute prothrombin time, kaolin clotting time, and some venom-based assays; thus, 58 of 286 (20%) centers failed to comply with this recommendation (Table 1).
In this proficiency testing exercise, two samples were distributed. Results were returned by a total of 291 centers. The false-positive error rate for the sample with mild FXI deficiency was 1.8% (five of 283 centers provided an interpretation of their results). Of these, four centers employed only APTT and DRVVT, with one center performing three different screening tests. Two centers reported a negative screen for the LA-positive sample. In both cases, their results were consistent with LA positivity, implying errors in interpretation. One of the two centers appeared to have transposed interpretations for the two samples. For centers employing APTT as an LA screening test, the guidelines recommend that the reagent activator be silica-based; 51 of 260 centers using APTT in this exercise (19.6%) employed reagents involving ellagic acid activation. Twenty of 260 (8%) centers employed a lupus-insensitive APTT reagent. The LA-positive sample was a high-potency LA, and the mean APTT ratio for this reagent was 2.54, which may account for the low false-negative rate. However, 13 centers employed an LA-insensitive APTT reagent in conjunction only with a DRVVT method, an approach that seems incompatible with the ISTH/SCC recommendations.
With respect to the DRVVT, 10 different methods or kits were used in the exercise described here. There is some ambiguity in the current guidelines. They recommend mixing tests and reporting ratios of test/pooled normal plasma for all assays, with a cut-off value based on the percentage correction of ratio algorithm. In contrast, however, the guidelines also state that integrated tests (e.g. test and confirm reagents) may not require mixing studies, and that the percentage correction of ratio may be useful. In this exercise, of 281 centers using DRVVT, 17 (6%) only reported a test/confirm ratio, with 125 (44%) reporting a normalized test/confirm ratio, and 52 (18%) determining the percentage correction of ratio. Two hundred and fifty-two centers (90%) reported a test/normal plasma ratio. In total, 277 (99%) centers performed a confirmatory test as part of their DRVVT assay, and 127 (45%) centers performed mixing studies for the DRVVT. In contrast, two of 260 (0.8%) centers reported phospholipid correction of their APTT, and 218/260 (83%) centers reported the results of APTT mixing studies. This indicates an approach whereby laboratories are compliant with the guidelines, in that mixing studies are generally performed on prolonged APTTs, and phospholipid correction is carried out with the DRVVT. However, the efficacy of this approach for an LA-positive sample manifesting only in the APTT assay has not been explored. The guidelines recommend that correction tests be carried out on the test for which the assay is prolonged.
Eighty-three of 295 centers employed locally prepared pooled plasma as a source of normal plasma, with the remainder using a commercial plasma. Of those preparing locally sourced plasma, pools of a minimum of five donors were employed.
With respect to reference ranges and interpretation of data, only 28 of 261 (11%) centers employed a cut-off value or reference range based on the 99th percentile; the majority (78%) employed mean + 2 standard deviations. One hundred and twenty-nine of 278 (46%) centers reported that they did not perform solid-phase assays as a routine part of their investigation. However, the questionnaire failed to ask whether laboratories referred these tests to other departments or institutions. Guidelines recommend that laboratories should provide interpretation of the results when reported to clinicians. Twenty-two centers (7%) did not provide an interpretative comment when reporting results.
Of the responses received in this exercise, 115 (39%) were from participating centers outside of the UK. In many aspects of practice, there were no appreciable differences between UK and non-UK centers. However, the following differences were noted: 0.129 m citrate anticoagulant was used by 3% of UK centers and 19% of non-UK centers; the 99th percentile was employed as a cut-off value in 3% of UK centers and 20% of non-UK centers; and interpretative data were provided by 96% of UK centers and 86% of non-UK centers.
We have demonstrated that compliance with the recent ISTH/SSC guidelines for LA testing is less than universal. We did not ask participants to identify awareness of the guidelines. However, our survey, distributed 6 months after publication of the guidelines, confirmed the findings in reports by other authors for surveys distributed prior to publication of these guidelines ; differences exist with respect to preanalytic variables, and particularly patterns of tests employed. The effects of many of these preanalytic variables cannot be assessed in proficiency testing exercises. Performance in LA testing exercises has improved since earlier reports, with a reduced false-positive rate for a factor-deficient sample. However, the challenge of more complex LA-negative samples (e.g. acquired FVIII inhibitors) and diagnosis of weak LA remains, and further standardization of methodology and approach may be required to improve diagnostic accuracy for this subgroup of patients.