Lupus anticoagulant detection: out of control?

Abstract

Expert guidelines indicate that normalised ratios are preferred to clotting times for lupus anticoagulant (LA) assays to mitigate analytical variation. We investigated the effects of deriving normalised ratios from the reference interval (RI) mean or different normal pooled plasmas (NPP). Screen, confirm and mixing tests for dilute Russell's viper venom time and APTT were converted to normalised ratios and interpreted for LA. Of 1000 clinical samples, 824 generated identical interpretations using RI mean or NPP-derived ratios and 57 identified LAs in one or both assays via either denominator. Separate RIs were applied for normalised ratios derived from the NPP or RI mean. Applying percentage correction index (PCI) to screen and confirm assays irrespective of screen elevation increased agreement to 92.5%. Two frozen and one lyophilised NPP were then used to derive ratios for 204 samples and 130 generated identical interpretations with all NPPs, 14 had overall interpretation parity and 19 overall agreement via PCI, giving 79.9% overall agreement. The results derived from each NPP were interpreted against RIs derived from RI means to reflect differences resulting from NPPs with clotting times dissimilar to RI means. Disparities were largely a function of closeness of NPP clotting times to test RI means and not owing to clotting factor level differences and likely related to manufacturing variables. Diagnostic benefit of normalised ratios can be maximised by matching NPP values to RI means. If RI mean is employed, and it likely requires re-establishing with new reagent batches.