Dilute Russel Viper Venom Time (DRVVT) Ratio Predicts the Identification of Triple Positive Antiphospholipid Antibody Profile

Introduction

The triple positive antiphospholipid (TP-aPL) antibody profile, i.e positive for lupus anticoagulant (LA), anti-cardiolipin antibodies (aCL) and anti-B2 glycoprotein I antibodies (aβ2GPI) have the highest risk of thrombotic events. Even in asymptomatic carriers, risk of a first thrombosis is increased especially in the presence of contributing venous thromboembolism (VTE) risk factors. Few studies have looked at the association of LA assays with odds of having a TP-aPL profile; double positivity (positive on aPTT and dRVVT arm) being a reported predictor(Da Silva Saraiva et al. 2017). Of the LA assays, the dRVVT has been shown to be more specific and robust with less variability compared to aPTT (activated partial thromboplastin time) or other systems. Thus, in this study, we analysed the association between the dRVVT screen/confirm ratio and risk of having a TP-aPL profile.

Methods

From our laboratory database, we identified all positive LA samples performed at our institution from June 2022 to December 2023. Only samples with paired aCL IgM/IgG and aβ2GPI IgM/IgG were included. Samples that were taken whilst patients were on anticoagulation were excluded. As per international guidelines, LA testing was performed with two tests based on different principles – dRVVT and aPTT(Devreese et al. 2020). For dRVVT, we used LA 1 Screening Reagent and LA 2 Confirmation Reagent (Siemens Healthcare, Marburg, Germany) on a Sysmex CS2500 analyzer (Sysmex Corporation, Kobe, Japan). Positivity was determined by the ratio of the dRVVT screen and confirmatory timings. The cut-off of 1.29 was established locally by mean (RI) ± 2SD in accordance to Clinical and Laboratory Standards Institute (CLSI) guidelines(Ledford-Kraemer 2014). Testing for aβ2GPI IgM/IgG was performed using chemiluminescent immunoassay (CLIA) and testing for aCL IgM/IgG performed using fluorescence enzyme immunoassay (FEIA). Manufacturer’s 99th centile cut-off was used to determine positivity(Devreese et al. 2020).

We analysed the association between DRVVT ratio with TP-aPL profile using linear regression and compared median ratios using Mann-Whitney U test. We subsequently performed ROC curve analysis to determine a potential DRVVT ratio cut-off that predicts risk of having TP-aPL profile. Optimal cut-off value was the maximum value for the Youden's index (J) which is obtained from the ROC curve analysis as the point that maximizes the sum of sensitivity plus specificity (J = sensitivity + specificity - 1). We subsequently validated this optimal cut-off on a separate cohort of positive LA samples that were identified from January 2024 to June 2024.

Results

A total of 244 positive LA samples met the study’s inclusion criteria and were analysed. 54 samples (22.1%) had TP-aPL profile. The median dRVVT ratio for the entire cohort was 1.49 (IQR 1.33 – 1.82). A higher dRVVT ratio was significantly associated with having a TP-aPL profile, OR 6.48 (95% CI: 2.21 to 18.94), p<0.001; i.e for every 0.1 increase in the dRVVT ratio, there was a 6.48 increased odds of the sample being TP-aPL. Samples with TP-aPL had a significantly higher median dRVVT then those without, 2.29 vs 1.40. The optimal cutoff value of DRVVT screen/confirm ratio according to ROC analysis was 1.81 (J = 0.74) with an area under the ROC curve (AUC) 0.922 (95% CI 0.881 - 0.953), p <0.001. Sensitivity and specificity were 83.3% and 91.1% respectively; positive predictive value (PPV) and negative predictive value (NPV) were 72.6% and 95.1% respectively. The overall accuracy was 89.3%. Using a dRVVT cut-off in predicting TP-aPL profile performed significantly better than using double positive LA assay (Sensitivity: 79.6%; Specificity: 71.1%; PPV: 43.9%; NPV 92.5%; Overall accuracy: 77.9%).

We subsequently tested the optimal dRVVT cut-off of 1.81 on a separate validation cohort. 123 LA positive samples from 1^st January 2024 to 30^th June 2024 that met the study’s inclusion criteria were included. 21 samples (17.1%) had the TP-aPL profile. Applying the dRVVT cut-off of 1.81, sensitivity and specificity for TP-aPL was 81.0% and 96.1% respectively. PPV and NPV were 81.0% and 96.1% respectively. The overall accuracy was 93.5%. These results were consistent with our development cohort.

Conclusion

In conclusion, our results demonstrate the utility of a dRVVT cut-off in predicting odds of TP-aPL profile vs non-TP aPL profile.