Recently, guidelines were published on the investigation and management of the antiphospholipid syndrome. (Greaves et al, 2000). These recommended that both lupus anticoagulant (LA) and solid-phase type assays must be used for the detection of antiphospholipid antibodies (aPLs) in certain patients because cases with LA but no anticardiolipin antibody (aCL), and vice versa, are well recognized, and diagnosis on the basis of just one type of assay may lead to false-negative aPL assessments.
We conducted an audit of all aPL results of requests from all sources over a 4 month period (January–April 2000). There were 521 requests over this period. Overall, both LA and aCL were tested in only 24% of cases (127 patients). Analysis by speciality showed a dual testing rate varying from 7% (requests from General Practitioners) to 63% from Paediatricians (however, the latter sample number was very small, only 3% of total requests). Even ‘informed’ users had a poor dual testing rate, being 41% from Haematology (representing 20% of total) and 32% from Immunology (10% of total). Of the 127 patients in whom both aCL and LA were tested, there were positive aCL results in 30 patients. Of these, 10 were LA-positive and 20 LA-negative. Of the LA-positive patients, one had a weakly positive aCL IgG (10–15), five moderately positive-IgM only (15–50) and four strongly positive (> 50), two of whom had IgG and IgM positivity, and two IgG only. Out of the 20 LA-negative patients, eight were weakly positive for aCL (four IgG, four IgM), 11 were moderately positive, seven of these being IgM only, and one strongly positive (IgM only). LA gave a positive predictive value for aCL of 33%.
In 172 cases, only LA was tested. Of these, 165 were negative and seven positive. Only aCL was tested in 222 samples. Of these, 192 were negative for IgG (190 for IgM), 10 weakly positive for IgG (15 IgM), 16 moderately positive for IgG (12 IgM) and four strongly positive for IgG (five IgM).
β2Glycoprotein 1 (β2GP1) was tested in 105 patients (at the discretion of the reporting Immunologist). In only 20 of these were both aCL and LA also tested. Ten of the 105 had a positive β2GP1 IgG (range 11–88·3). Of these, three were tested for LA. One was positive and two negative. All 10 were tested for aCL, eight being positive. Four of the 10 were also positive for β2GP1 IgM. Finally, in four of the 20 samples, the β2GP1 was an isolated abnormality.
We concluded that the dual testing rate was poor, even in informed users. When data for all tests were available, strongly positive aCL and LA did not always coincide. Both tests were required to identify all abnormal patients. Hence, we concur that both tests are required as stated in the APL guidelines. Systems need to be in place to ensure dual testing for aCL and LA. This is complicated by the need to send different samples to two different departments.
The significance of an isolated abnormal β2GP1 is uncertain, but may be caused by species-specific antibodies that fail to bind to bovine β2GP1 in the aCL assay. When all the data were available, the β2GP1 increased the positive predictive value of LA to 75% in this small sample. However, it should not replace LA or aCL.
Finally, there is debate about the importance of aCL isotype. IgG antibodies may be more significant clinically, although IgM aCL appears to be associated with thrombotic events and miscarriages in some cases (Greaves et al, 2000). In our study, of all LA-positive patients, six had only IgM aCLs – these would have been missed if only IgG aCLs were tested.