Revisiting the Phadia/EliA cut-off values for anticardiolipin and anti-β2-glycoprotein I antibodies: a systematic...

Background Phadia/EliA fluorescence enzyme immunoassays are widely used automated assays for anticardiolipin (aCL) and anti-β2-glycoprotein I (aβ2GPI) antibodies. To date, cut-off values for these assays have not been evaluated systematically and the evidence behind manufacturer's recommended cut-off values is not clear. Objective To determine Phadia/EliA cut-off values for antiphospholipid antibodies (aPL) according to the procedures suggested by guidelines. Methods A total of 266 blood donors (135 females and 131 males) were included. The pre-handling and analysis of the samples were performed according to the International Society on Thrombosis and Hemostasis (ISTH) guideline for solid phase aPL assays. Cut-off values and corresponding 90% confidence intervals (CI) for each antibody were established and outliers were handled according to the Clinical and Laboratory Standards Institute (CLSI) guideline for reference intervals. Samples from 377 consecutive patients, referred to our thrombophilia center with evidence of thrombosis or pregnancy morbidity were included for aPL testing. Results The in-house 99th (97.5th) percentile cut-off values were 11 (8.7), 12 (6.9) 8.5 (5.0) AU/mL for aβ2GPI IgG, IgM and IgA, and 21 (13) GPL-U/mL and 41 (25) MPL-U/mL for aCL IgG and IgM, respectively. The prevalence of positive results (%) defined by these cut-off values in patients with evidence of thrombosis or pregnancy morbidity was 9.5 (12.2), 1.6 (2.9), and 7.0 (9.9), and 0.8 (3.8) for aβ2GPI IgG, IgM, and aCL IgG and IgM respectively. The use of in-house 99th percentile cut-off values compared to the manufacturer suggested cut-off values resulted in 1 and 39 fewer samples for aβ2GPI and aCL to be classified as positive for aPL, respectively. Conclusions We present Phadia/EliA cut-off values with 90% CI for aPL determined systematically according to the ISTH and CLSI guidelines. These values are different from values previously determined, suggesting variation of aPLs in different populations. Our findings indicate the need for each laboratory to determine/validate assay specific cut-off values for aPL.