A-146 Classic vs. Integrated Testing Algorithms for Lupus Anticoagulant: How to Proceed?

Abstract Background Lupus anticoagulant (LA) testing guidelines recommend the use of two paired phospholipid-dependent, clot-based assays measured in seconds, a dilute Russel Viper Venom Time (dRVVT) and a Partial Thromboplastin Time (PTT)-based assay, with two steps: a screen step (low phospholipid) and a confirm step (high phospholipid). In current recommended algorithms, the confirm step is only performed reflexively if the screen step is prolonged (“classic algorithm”). However, some manufacturers’ FDA-cleared instructions for use (MIFU) describe performing screen and confirm steps simultaneously (“integrated algorithm”). Thus, performing the screen and confirm steps via the classic algorithm would classify this as a laboratory developed test (LDT). Given the pending FDA changes in the United States, many laboratories are attempting to minimize LDT’s. Limited data exist comparing the classic and integrated algorithms, especially on results that would only be considered positive using the integrated algorithm due to normal screen times. We sought to compare the results of these two algorithms, using the silica clot time (SCT) PTT-based LA assay (Werfen, Bedford, MA) at two institutions. Methods We performed a 6-month, retrospective review at two regional laboratories using the integrated SCT method to identify samples with “normal SCT screen and positive SCT Test Ratio” (NSTR+) following the Werfen MIFU. For NSTR+ samples, we also collected dRVVT and immunoassay results [anticardiolipin (aCL) and beta-2 glycoprotein-1 (B2GP1) antibody, IgG and IgM]. Results LA testing was performed on 2,708 samples during a 6-month time period. Overall, there were 410 LA-positive specimens by clot-based assay dRVVT and/or SCT; 112 (27.3% of total LA-positive) were NSTR+ by integrated algorithm. For these NSTR+ samples, mean patient age was 48.6 (range 15-90) years with 86 (77%) female. Seventy-eight NSTR+ samples (70% of NSTR+, 19% of total LA-positive) were negative by dRVVT and would be reported as LA-negative by the classic algorithm. Forty NSTR+ samples (36% of NSTR+, 9.8 % of total LA-positive) showed positivity by 1) dRVVT with or without immunoassay positivity (34 samples), or 2) immunoassay only (4 samples either aCL or B2GP1 antibody positive; 2 samples both aCL and B2GP1 antibody positive). The integrated algorithm identified significantly more positive samples than the classic algorithm (Fisher’s exact test p<0.00001). Conclusion By detecting phospholipid dependence in samples with normal screen steps (NSTR+), the integrated algorithm demonstrates increased sensitivity as compared to the classic algorithm. Thirty-six percent of the NSTR+ samples showed positivity by an alternative testing method, suggesting that the SCT classic algorithm is missing at least a subset of true positive samples. While the clinical significance of this is unclear, additional guidance will be necessary for laboratories and clinicians for the investigation of LA using paired assay systems. Further investigation, including clinical correlation and subsequent confirmation of positivity, will be necessary.